Vertex AI release notes

<!doctype html> <html lang="en" dir="ltr"> <head> <meta name="google-signin-client-id" content="721724668570-nbkv1cfusk7kk4eni4pjvepaus73b13t.apps.googleusercontent.com"> <meta name="google-signin-scope" content="profile email https://www.googleapis.com/auth/developerprofiles https://www.googleapis.com/auth/developerprofiles.award https://www.googleapis.com/auth/cloud-platform https://www.googleapis.com/auth/webhistory"> <meta property="og:site_name" content="Google Cloud"> <meta property="og:type" content="website"><meta name="theme-color" content="#039be5"><meta charset="utf-8"> <meta content="IE=Edge" http-equiv="X-UA-Compatible"> <meta name="viewport" content="width=device-width, initial-scale=1"> <link rel="manifest" href="/_pwa/cloud/manifest.json" crossorigin="use-credentials"> <link rel="preconnect" href="//www.gstatic.com" crossorigin> <link rel="preconnect" href="//fonts.gstatic.com" crossorigin> <link rel="preconnect" href="//fonts.googleapis.com" crossorigin> <link rel="preconnect" href="//apis.google.com" crossorigin> <link rel="preconnect" href="//www.google-analytics.com" crossorigin><link rel="stylesheet" href="//fonts.googleapis.com/css?family=Google+Sans:400,500,700|Google+Sans+Text:400,400italic,500,500italic,700,700italic|Roboto:400,400italic,500,500italic,700,700italic|Roboto+Mono:400,500,700&display=swap"> <link rel="stylesheet" href="//fonts.googleapis.com/css2?family=Material+Icons&family=Material+Symbols+Outlined&display=block"><link rel="stylesheet" href="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/css/app.css"> <link rel="shortcut icon" href="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/favicons/onecloud/favicon.ico"> <link rel="apple-touch-icon" href="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/favicons/onecloud/super_cloud.png"><link rel="canonical" href="https://cloud.google.com/vertex-ai/docs/release-notes"><link rel="search" type="application/opensearchdescription+xml" title="Google Cloud" href="https://cloud.google.com/s/opensearch.xml"> <link rel="alternate" hreflang="en" href="https://cloud.google.com/vertex-ai/docs/release-notes" /><link rel="alternate" hreflang="x-default" href="https://cloud.google.com/vertex-ai/docs/release-notes" /><title>Vertex AI release notes  |  Google Cloud</title> <meta property="og:title" content="Vertex AI release notes  |  Google Cloud"><meta property="og:url" content="https://cloud.google.com/vertex-ai/docs/release-notes"><meta property="og:image" content="https://cloud.google.com/_static/cloud/images/social-icon-google-cloud-1200-630.png"> <meta property="og:image:width" content="1200"> <meta property="og:image:height" content="630"><meta property="og:locale" content="en"><meta name="twitter:card" content="summary_large_image"><script type="application/ld+json"> { "@context": "https://schema.org", "@type": "Article", "headline": "Vertex AI release notes" } </script><script type="application/ld+json"> { "@context": "https://schema.org", "@type": "BreadcrumbList", "itemListElement": [{ "@type": "ListItem", "position": 1, "name": "Vertex AI", "item": "https://cloud.google.com/vertex-ai" },{ "@type": "ListItem", "position": 2, "name": "Documentation", "item": "https://cloud.google.com/vertex-ai/docs" },{ "@type": "ListItem", "position": 3, "name": "Vertex AI release notes", "item": "https://cloud.google.com/vertex-ai/docs/release-notes" }] } </script> <link rel="alternate" type="application/atom+xml" href="https://cloud.google.com/feeds/vertex-ai-product-group-release-notes.xml"/></p> <p></p> <link rel="alternate" type="application/atom+xml" href="https://cloud.google.com/feeds/vertex-ai-product-group-release-notes.xml"/> <meta name="gtm_var" data-key="docType" data-value="release-notes"> <link rel="stylesheet" href="/extras.css"></head> <body class="" template="page" theme="cloud-theme" type="article" layout="docs" free-trial display-toc pending> <devsite-progress type="indeterminate" id="app-progress"></devsite-progress> <a href="#main-content" class="skip-link button"> Skip to main content </a> <section class="devsite-wrapper"> <devsite-cookie-notification-bar></devsite-cookie-notification-bar><cloudx-track userCountry="SG"></cloudx-track> <cloudx-utils-init></cloudx-utils-init> <devsite-header keep-tabs-visible> <div class="devsite-header--inner nocontent"> <div class="devsite-top-logo-row-wrapper-wrapper"> <div class="devsite-top-logo-row-wrapper"> <div class="devsite-top-logo-row"> <button type="button" id="devsite-hamburger-menu" class="devsite-header-icon-button button-flat material-icons gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Navigation menu button" visually-hidden aria-label="Open menu"> </button> <div class="devsite-product-name-wrapper"> <a href="/" class="devsite-site-logo-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Site logo" track-type="globalNav" track-name="googleCloud" track-metadata-position="nav" track-metadata-eventDetail="nav"> <picture> <img src="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/cloud-logo.svg" class="devsite-site-logo" alt="Google Cloud"> </picture> </a> <span class="devsite-product-name"> <ul class="devsite-breadcrumb-list" > <li class="devsite-breadcrumb-item devsite-has-google-wordmark"> </li> </ul> </span> </div> <div class="devsite-top-logo-row-middle"> <div class="devsite-header-upper-tabs"> <cloudx-tabs-nav class="upper-tabs"> <nav class="devsite-tabs-wrapper" aria-label="Upper tabs"> <tab class="devsite-active"> <a href="https://cloud.google.com/docs" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/docs" track-type="nav" track-metadata-position="nav - docs-home" track-metadata-module="primary nav" aria-label="Documentation, selected" data-category="Site-Wide Custom Events" data-label="Tab: Documentation" track-name="docs-home" track-link-column-type="single-column" > Documentation </a> </tab> <tab class="devsite-dropdown devsite-clickable "> <a href="https://cloud.google.com/docs/tech-area-overviews" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/docs/tech-area-overviews" track-type="nav" track-metadata-position="nav - technology-areas" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Technology areas" track-name="technology-areas" track-link-column-type="single-column" > Technology areas </a> <button aria-haspopup="menu" aria-expanded="false" aria-label="Dropdown menu for Technology areas" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/tech-area-overviews" track-metadata-position="nav - technology-areas" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Technology areas" track-name="technology-areas" track-link-column-type="single-column" class="devsite-tabs-dropdown-toggle devsite-icon devsite-icon-arrow-drop-down"></button> <div class="devsite-tabs-dropdown" role="menu" aria-label="submenu" hidden> <button class="devsite-tabs-close-button material-icons button-flat gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Close dropdown menu" aria-label="Close dropdown menu" track-type="nav" track-name="close" track-metadata-eventdetail="#" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav">close</button> <div class="devsite-tabs-dropdown-content"> <div class="devsite-tabs-dropdown-column "> <ul class="devsite-tabs-dropdown-section "> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/ai-ml" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/ai-ml" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> AI and ML </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/application-development" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/application-development" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Application development </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/application-hosting" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/application-hosting" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Application hosting </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/compute-area" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/compute-area" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Compute </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/data" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/data" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Data analytics and pipelines </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/databases" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/databases" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Databases </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/dhm-cloud" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/dhm-cloud" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Distributed, hybrid, and multicloud </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/generative-ai" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/generative-ai" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Generative AI </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/industry" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/industry" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Industry solutions </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/networking" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/networking" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Networking </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/observability" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/observability" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Observability and monitoring </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/security" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/security" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Security </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/storage" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/storage" track-metadata-position="nav - technology-areas" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Storage </div> </a> </li> </ul> </div> </div> </div> </tab> <tab class="devsite-dropdown devsite-clickable "> <a href="https://cloud.google.com/docs/cross-product-overviews" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/docs/cross-product-overviews" track-type="nav" track-metadata-position="nav - crossproduct" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Cross-product tools" track-name="crossproduct" track-link-column-type="single-column" > Cross-product tools </a> <button aria-haspopup="menu" aria-expanded="false" aria-label="Dropdown menu for Cross-product tools" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/cross-product-overviews" track-metadata-position="nav - crossproduct" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Cross-product tools" track-name="crossproduct" track-link-column-type="single-column" class="devsite-tabs-dropdown-toggle devsite-icon devsite-icon-arrow-drop-down"></button> <div class="devsite-tabs-dropdown" role="menu" aria-label="submenu" hidden> <button class="devsite-tabs-close-button material-icons button-flat gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Close dropdown menu" aria-label="Close dropdown menu" track-type="nav" track-name="close" track-metadata-eventdetail="#" track-metadata-position="nav - crossproduct" track-metadata-module="tertiary nav">close</button> <div class="devsite-tabs-dropdown-content"> <div class="devsite-tabs-dropdown-column "> <ul class="devsite-tabs-dropdown-section "> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/access-resources" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/access-resources" track-metadata-position="nav - crossproduct" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Access and resources management </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/costs-usage" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/costs-usage" track-metadata-position="nav - crossproduct" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Costs and usage management </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/devtools" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/devtools" track-metadata-position="nav - crossproduct" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud SDK, languages, frameworks, and tools </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/iac" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/iac" track-metadata-position="nav - crossproduct" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Infrastructure as code </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/docs/migration" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/docs/migration" track-metadata-position="nav - crossproduct" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Migration </div> </a> </li> </ul> </div> </div> </div> </tab> <tab class="devsite-dropdown devsite-clickable "> <a href="https://cloud.google.com/" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/" track-type="nav" track-metadata-position="nav - related-sites" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Related sites" track-name="related-sites" track-link-column-type="single-column" > Related sites </a> <button aria-haspopup="menu" aria-expanded="false" aria-label="Dropdown menu for Related sites" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/" track-metadata-position="nav - related-sites" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Related sites" track-name="related-sites" track-link-column-type="single-column" class="devsite-tabs-dropdown-toggle devsite-icon devsite-icon-arrow-drop-down"></button> <div class="devsite-tabs-dropdown" role="menu" aria-label="submenu" hidden> <button class="devsite-tabs-close-button material-icons button-flat gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Close dropdown menu" aria-label="Close dropdown menu" track-type="nav" track-name="close" track-metadata-eventdetail="#" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav">close</button> <div class="devsite-tabs-dropdown-content"> <div class="devsite-tabs-dropdown-column "> <ul class="devsite-tabs-dropdown-section "> <li class="devsite-nav-item"> <a href="https://cloud.google.com/" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Home </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/free" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/free" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Free Trial and Free Tier </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/architecture" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/architecture" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Architecture Center </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/blog" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/blog" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Blog </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/contact" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/contact" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Contact Sales </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/developers" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/developers" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Developer Center </div> </a> </li> <li class="devsite-nav-item"> <a href="https://developers.google.com/" track-type="nav" track-metadata-eventdetail="https://developers.google.com/" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Developer Center </div> </a> </li> <li class="devsite-nav-item"> <a href="https://console.cloud.google.com/marketplace" track-type="nav" track-metadata-eventdetail="https://console.cloud.google.com/marketplace" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Marketplace </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/marketplace/docs" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/marketplace/docs" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Marketplace Documentation </div> </a> </li> <li class="devsite-nav-item"> <a href="https://www.cloudskillsboost.google/paths" track-type="nav" track-metadata-eventdetail="https://www.cloudskillsboost.google/paths" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Skills Boost </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/solutions" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/solutions" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Solution Center </div> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/support-hub" track-type="nav" track-metadata-eventdetail="https://cloud.google.com/support-hub" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Support </div> </a> </li> <li class="devsite-nav-item"> <a href="https://www.youtube.com/@googlecloudtech" track-type="nav" track-metadata-eventdetail="https://www.youtube.com/@googlecloudtech" track-metadata-position="nav - related-sites" track-metadata-module="tertiary nav" tooltip > <div class="devsite-nav-item-title"> Google Cloud Tech Youtube Channel </div> </a> </li> </ul> </div> </div> </div> </tab> </nav> </cloudx-tabs-nav> </div> <devsite-search enable-signin enable-search enable-suggestions project-name="Vertex AI" tenant-name="Google Cloud" project-scope="/vertex-ai/docs" url-scoped="https://cloud.google.com/s/results/vertex-ai/docs" > <form class="devsite-search-form" action="https://cloud.google.com/s/results" method="GET"> <div class="devsite-search-container"> <button type="button" search-open class="devsite-search-button devsite-header-icon-button button-flat material-icons" aria-label="Open search"></button> <div class="devsite-searchbox"> <input aria-activedescendant="" aria-autocomplete="list" aria-label="Search" aria-expanded="false" aria-haspopup="listbox" autocomplete="off" class="devsite-search-field devsite-search-query" name="q" placeholder="Search" role="combobox" type="text" value="" > <div class="devsite-search-image material-icons" aria-hidden="true"> </div> <div class="devsite-search-shortcut-icon-container" aria-hidden="true"> <kbd class="devsite-search-shortcut-icon">/</kbd> </div> </div> </div> </form> <button type="button" search-close class="devsite-search-button devsite-header-icon-button button-flat material-icons" aria-label="Close search"></button> </devsite-search> </div> <devsite-language-selector> <ul role="presentation"> <li role="presentation"> <a role="menuitem" lang="en" >English</a> </li> <li role="presentation"> <a role="menuitem" lang="de" >Deutsch</a> </li> <li role="presentation"> <a role="menuitem" lang="es_419" >Español – América Latina</a> </li> <li role="presentation"> <a role="menuitem" lang="fr" >Français</a> </li> <li role="presentation"> <a role="menuitem" lang="pt_br" >Português – Brasil</a> </li> <li role="presentation"> <a role="menuitem" lang="zh_cn" >中文 – 简体</a> </li> <li role="presentation"> <a role="menuitem" lang="ja" >日本語</a> </li> <li role="presentation"> <a role="menuitem" lang="ko" >한국어</a> </li> </ul> </devsite-language-selector> <devsite-user enable-profiles fp-auth id="devsite-user"> <span class="button devsite-top-button" aria-hidden="true" visually-hidden>Sign in</span> </devsite-user> </div> </div> </div> <div class="devsite-collapsible-section "> <div class="devsite-header-background"> <div class="devsite-product-id-row" hidden> <div class="devsite-product-description-row"> </div> </div> <div class="devsite-doc-set-nav-row"> <ul class="devsite-breadcrumb-list" > <li class="devsite-breadcrumb-item "> <a href="https://cloud.google.com/vertex-ai" class="devsite-breadcrumb-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Lower Header" data-value="1" track-type="globalNav" track-name="breadcrumb" track-metadata-position="1" track-metadata-eventdetail="Vertex AI" > Vertex AI </a> </li> </ul> <cloudx-tabs-nav class="lower-tabs"> <nav class="devsite-tabs-wrapper" aria-label="Lower tabs"> <tab > <a href="https://cloud.google.com/vertex-ai/docs" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/vertex-ai/docs" track-type="nav" track-metadata-position="nav - guides" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Guides" track-name="guides" > Guides </a> </tab> <tab > <a href="https://cloud.google.com/vertex-ai/docs/reference" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/vertex-ai/docs/reference" track-type="nav" track-metadata-position="nav - reference" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Reference" track-name="reference" > Reference </a> </tab> <tab > <a href="https://cloud.google.com/vertex-ai/docs/samples" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/vertex-ai/docs/samples" track-type="nav" track-metadata-position="nav - samples" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Samples" track-name="samples" > Samples </a> </tab> <tab > <a href="https://cloud.google.com/vertex-ai/docs/support/getting-support" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/vertex-ai/docs/support/getting-support" track-type="nav" track-metadata-position="nav - support" track-metadata-module="primary nav" data-category="Site-Wide Custom Events" data-label="Tab: Support" track-name="support" > Support </a> </tab> <tab class="devsite-active"> <a href="https://cloud.google.com/vertex-ai/docs/release-notes" class="devsite-tabs-content gc-analytics-event " track-metadata-eventdetail="https://cloud.google.com/vertex-ai/docs/release-notes" track-type="nav" track-metadata-position="nav - resources" track-metadata-module="primary nav" aria-label="Resources, selected" data-category="Site-Wide Custom Events" data-label="Tab: Resources" track-name="resources" > Resources </a> </tab> </nav> </cloudx-tabs-nav> <div class="devsite-product-button-row"> <a href="https://cloud.google.com/contact" class="cta-button-secondary button " track-type="contact" track-metadata-position="nav" data-overflow-wrapper="tab" data-overflow="devsite-tabs-wrapper" track-metadata-eventDetail="nav" track-name="sales" data-overflow-container="left" >Contact Us</a> <a href="//console.cloud.google.com/freetrial" class="cloud-free-trial-button cta-button-primary button-primary button cloud-button cloud-button--primary " data-overflow-container="right" track-type="freeTrial" track-metadata-position="nav" track-metadata-eventDetail="nav" data-overflow-class="devsite-header-link devsite-top-button button cloud-free-trial-button cloud-free-trial-enabled cloud-button cloud-button--primary" data-overflow="devsite-top-logo-row" referrerpolicy="no-referrer-when-downgrade" track-name="gcpCta" >Start free</a> </div> </div> </div> </div> </div> </devsite-header> <devsite-book-nav scrollbars > <div class="devsite-book-nav-filter" > <span class="filter-list-icon material-icons" aria-hidden="true"></span> <input type="text" placeholder="Filter" aria-label="Type to filter" role="searchbox"> <span class="filter-clear-button hidden" data-title="Clear filter" aria-label="Clear filter" role="button" tabindex="0"></span> </div> <nav class="devsite-book-nav devsite-nav nocontent" aria-label="Side menu"> <div class="devsite-mobile-header"> <button type="button" id="devsite-close-nav" class="devsite-header-icon-button button-flat material-icons gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Close navigation" aria-label="Close navigation"> </button> <div class="devsite-product-name-wrapper"> <a href="/" class="devsite-site-logo-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Site logo" track-type="globalNav" track-name="googleCloud" track-metadata-position="nav" track-metadata-eventDetail="nav"> <picture> <img src="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/cloud-logo.svg" class="devsite-site-logo" alt="Google Cloud"> </picture> </a> <span class="devsite-product-name"> <ul class="devsite-breadcrumb-list" > <li class="devsite-breadcrumb-item devsite-has-google-wordmark"> </li> </ul> </span> </div> </div> <div class="devsite-book-nav-wrapper"> <div class="devsite-mobile-nav-top"> <ul class="devsite-nav-list"> <li class="devsite-nav-item"> <a href="/docs" class="devsite-nav-title gc-analytics-event devsite-nav-active" data-category="Site-Wide Custom Events" data-label="Tab: Documentation" track-name="docs-home" track-link-column-type="single-column" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Documentation" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Documentation </span> </a> <ul class="devsite-nav-responsive-tabs"> <li class="devsite-nav-item"> <a href="/vertex-ai/docs" class="devsite-nav-title gc-analytics-event devsite-nav-has-children " data-category="Site-Wide Custom Events" data-label="Tab: Guides" track-name="guides" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Guides" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Guides </span> <span class="devsite-nav-icon material-icons" data-icon="forward" > </span> </a> </li> <li class="devsite-nav-item"> <a href="/vertex-ai/docs/reference" class="devsite-nav-title gc-analytics-event devsite-nav-has-children " data-category="Site-Wide Custom Events" data-label="Tab: Reference" track-name="reference" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Reference" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Reference </span> <span class="devsite-nav-icon material-icons" data-icon="forward" > </span> </a> </li> <li class="devsite-nav-item"> <a href="/vertex-ai/docs/samples" class="devsite-nav-title gc-analytics-event devsite-nav-has-children " data-category="Site-Wide Custom Events" data-label="Tab: Samples" track-name="samples" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Samples" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Samples </span> <span class="devsite-nav-icon material-icons" data-icon="forward" > </span> </a> </li> <li class="devsite-nav-item"> <a href="/vertex-ai/docs/support/getting-support" class="devsite-nav-title gc-analytics-event devsite-nav-has-children " data-category="Site-Wide Custom Events" data-label="Tab: Support" track-name="support" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Support" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Support </span> <span class="devsite-nav-icon material-icons" data-icon="forward" > </span> </a> </li> <li class="devsite-nav-item"> <a href="/vertex-ai/docs/release-notes" class="devsite-nav-title gc-analytics-event devsite-nav-has-children devsite-nav-active" data-category="Site-Wide Custom Events" data-label="Tab: Resources" track-name="resources" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Resources" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip menu="_book"> Resources </span> <span class="devsite-nav-icon material-icons" data-icon="forward" menu="_book"> </span> </a> </li> </ul> </li> <li class="devsite-nav-item"> <a href="/docs/tech-area-overviews" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Tab: Technology areas" track-name="technology-areas" track-link-column-type="single-column" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Technology areas" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Technology areas </span> </a> <ul class="devsite-nav-responsive-tabs devsite-nav-has-menu "> <li class="devsite-nav-item"> <span class="devsite-nav-title" tooltip data-category="Site-Wide Custom Events" data-label="Tab: Technology areas" track-name="technology-areas" track-link-column-type="single-column" > <span class="devsite-nav-text" tooltip menu="Technology areas"> More </span> <span class="devsite-nav-icon material-icons" data-icon="forward" menu="Technology areas"> </span> </span> </li> </ul> </li> <li class="devsite-nav-item"> <a href="/docs/cross-product-overviews" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Tab: Cross-product tools" track-name="crossproduct" track-link-column-type="single-column" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Cross-product tools" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Cross-product tools </span> </a> <ul class="devsite-nav-responsive-tabs devsite-nav-has-menu "> <li class="devsite-nav-item"> <span class="devsite-nav-title" tooltip data-category="Site-Wide Custom Events" data-label="Tab: Cross-product tools" track-name="crossproduct" track-link-column-type="single-column" > <span class="devsite-nav-text" tooltip menu="Cross-product tools"> More </span> <span class="devsite-nav-icon material-icons" data-icon="forward" menu="Cross-product tools"> </span> </span> </li> </ul> </li> <li class="devsite-nav-item"> <a href="/" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Tab: Related sites" track-name="related-sites" track-link-column-type="single-column" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Related sites" track-type="globalNav" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Related sites </span> </a> <ul class="devsite-nav-responsive-tabs devsite-nav-has-menu "> <li class="devsite-nav-item"> <span class="devsite-nav-title" tooltip data-category="Site-Wide Custom Events" data-label="Tab: Related sites" track-name="related-sites" track-link-column-type="single-column" > <span class="devsite-nav-text" tooltip menu="Related sites"> More </span> <span class="devsite-nav-icon material-icons" data-icon="forward" menu="Related sites"> </span> </span> </li> </ul> </li> <li class="devsite-nav-item"> <a href="//console.cloud.google.com/" class="devsite-nav-title gc-analytics-event " referrerpolicy="no-referrer-when-downgrade" track-metadata-eventDetail="nav" track-type="globalNav" track-name="console" track-metadata-position="nav" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Console" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Console </span> </a> </li> <li class="devsite-nav-item"> <a href="/contact" class="cta-button-secondary button" track-type="contact" track-metadata-position="nav" data-overflow-wrapper="tab" data-overflow="devsite-tabs-wrapper" track-metadata-eventDetail="nav" track-name="sales" data-overflow-container="left" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Contact Us" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Contact Us </span> </a> </li> <li class="devsite-nav-item"> <a href="//console.cloud.google.com/freetrial" class="cloud-free-trial-button cta-button-primary button-primary button cloud-button cloud-button--primary" data-overflow-container="right" track-type="freeTrial" track-metadata-position="nav" track-metadata-eventDetail="nav" data-overflow-class="devsite-header-link devsite-top-button button cloud-free-trial-button cloud-free-trial-enabled cloud-button cloud-button--primary" data-overflow="devsite-top-logo-row" referrerpolicy="no-referrer-when-downgrade" track-name="gcpCta" data-category="Site-Wide Custom Events" data-label="Responsive Tab: Start free" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Start free </span> </a> </li> </ul> </div> <div class="devsite-mobile-nav-bottom"> <ul class="devsite-nav-list" menu="_book"> <li class="devsite-nav-item devsite-nav-heading"><div class="devsite-nav-title devsite-nav-title-no-path"> <span class="devsite-nav-text" tooltip>Vertex AI</span> </div></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/release-notes" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/release-notes" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/release-notes" ><span class="devsite-nav-text" tooltip>Release notes</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/glossary" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/glossary" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/glossary" ><span class="devsite-nav-text" tooltip>Glossary</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/pricing" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/pricing" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/pricing" ><span class="devsite-nav-text" tooltip>Pricing</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/quotas" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/quotas" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/quotas" ><span class="devsite-nav-text" tooltip>Quotas and limits</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/general/locations" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/general/locations" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/general/locations" ><span class="devsite-nav-text" tooltip>Locations</span></a></li> <li class="devsite-nav-item devsite-nav-expandable"><div class="devsite-expandable-nav"> <a class="devsite-nav-toggle" aria-hidden="true"></a><div class="devsite-nav-title devsite-nav-title-no-path" tabindex="0" role="button"> <span class="devsite-nav-text" tooltip>Support policy</span> </div><ul class="devsite-nav-section"><li class="devsite-nav-item"><a href="/vertex-ai/docs/shared-responsibility" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/shared-responsibility" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/shared-responsibility" ><span class="devsite-nav-text" tooltip>Shared responsibility</span></a></li><li class="devsite-nav-item"><a href="/vertex-ai/docs/framework-support-policy" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/framework-support-policy" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/framework-support-policy" ><span class="devsite-nav-text" tooltip>Vertex AI framework support policy</span></a></li><li class="devsite-nav-item"><a href="/vertex-ai/docs/supported-frameworks-list" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/supported-frameworks-list" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/supported-frameworks-list" ><span class="devsite-nav-text" tooltip>Supported frameworks list</span></a></li></ul></div></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/security-bulletins" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/security-bulletins" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/security-bulletins" ><span class="devsite-nav-text" tooltip>Security bulletins</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/sla" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/sla" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/sla" ><span class="devsite-nav-text" tooltip>Service Level Agreement</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/general/supported-languages" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/general/supported-languages" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/general/supported-languages" ><span class="devsite-nav-text" tooltip>Supported languages for AutoML text models</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/general/vertexai-compliance" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/general/vertexai-compliance" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/general/vertexai-compliance" ><span class="devsite-nav-text" tooltip>Certifications</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/general/vertexai-security-controls" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/general/vertexai-security-controls" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/general/vertexai-security-controls" ><span class="devsite-nav-text" tooltip>Security controls</span></a></li> <li class="devsite-nav-item"><a href="/vertex-ai/docs/deprecations" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/deprecations" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/deprecations" ><span class="devsite-nav-text" tooltip>Vertex AI deprecations</span></a></li> <li class="devsite-nav-item devsite-nav-expandable"><div class="devsite-expandable-nav"> <a class="devsite-nav-toggle" aria-hidden="true"></a><div class="devsite-nav-title devsite-nav-title-no-path" tabindex="0" role="button"> <span class="devsite-nav-text" tooltip>AI Agent Ecosystem partners</span> </div><ul class="devsite-nav-section"><li class="devsite-nav-item"><a href="/vertex-ai/docs/ai-agent-ecosystem-overview" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/ai-agent-ecosystem-overview" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/ai-agent-ecosystem-overview" ><span class="devsite-nav-text" tooltip>Overview</span></a></li><li class="devsite-nav-item"><a href="/vertex-ai/docs/ai-agent-ecosystem-partners" class="devsite-nav-title gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Book nav link, pathname: /vertex-ai/docs/ai-agent-ecosystem-partners" track-type="bookNav" track-name="click" track-metadata-eventdetail="/vertex-ai/docs/ai-agent-ecosystem-partners" ><span class="devsite-nav-text" tooltip>AI Agent Ecosystem partners</span></a></li></ul></div></li> </ul> <ul class="devsite-nav-list" menu="Technology areas" aria-label="Side menu" hidden> <li class="devsite-nav-item"> <a href="/docs/ai-ml" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: AI and ML" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > AI and ML </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/application-development" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Application development" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Application development </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/application-hosting" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Application hosting" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Application hosting </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/compute-area" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Compute" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Compute </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/data" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Data analytics and pipelines" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Data analytics and pipelines </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/databases" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Databases" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Databases </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/dhm-cloud" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Distributed, hybrid, and multicloud" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Distributed, hybrid, and multicloud </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/generative-ai" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Generative AI" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Generative AI </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/industry" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Industry solutions" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Industry solutions </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/networking" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Networking" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Networking </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/observability" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Observability and monitoring" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Observability and monitoring </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/security" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Security" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Security </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/storage" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Storage" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Storage </span> </a> </li> </ul> <ul class="devsite-nav-list" menu="Cross-product tools" aria-label="Side menu" hidden> <li class="devsite-nav-item"> <a href="/docs/access-resources" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Access and resources management" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Access and resources management </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/costs-usage" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Costs and usage management" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Costs and usage management </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/devtools" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud SDK, languages, frameworks, and tools" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud SDK, languages, frameworks, and tools </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/iac" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Infrastructure as code" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Infrastructure as code </span> </a> </li> <li class="devsite-nav-item"> <a href="/docs/migration" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Migration" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Migration </span> </a> </li> </ul> <ul class="devsite-nav-list" menu="Related sites" aria-label="Side menu" hidden> <li class="devsite-nav-item"> <a href="/" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Home" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Home </span> </a> </li> <li class="devsite-nav-item"> <a href="/free" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Free Trial and Free Tier" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Free Trial and Free Tier </span> </a> </li> <li class="devsite-nav-item"> <a href="/architecture" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Architecture Center" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Architecture Center </span> </a> </li> <li class="devsite-nav-item"> <a href="https://cloud.google.com/blog" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Blog" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Blog </span> </a> </li> <li class="devsite-nav-item"> <a href="/contact" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Contact Sales" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Contact Sales </span> </a> </li> <li class="devsite-nav-item"> <a href="/developers" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Developer Center" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Developer Center </span> </a> </li> <li class="devsite-nav-item"> <a href="https://developers.google.com/" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Developer Center" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Developer Center </span> </a> </li> <li class="devsite-nav-item"> <a href="https://console.cloud.google.com/marketplace" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Marketplace" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Marketplace </span> </a> </li> <li class="devsite-nav-item"> <a href="/marketplace/docs" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Marketplace Documentation" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Marketplace Documentation </span> </a> </li> <li class="devsite-nav-item"> <a href="https://www.cloudskillsboost.google/paths" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Skills Boost" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Skills Boost </span> </a> </li> <li class="devsite-nav-item"> <a href="/solutions" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Solution Center" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Solution Center </span> </a> </li> <li class="devsite-nav-item"> <a href="/support-hub" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Support" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Support </span> </a> </li> <li class="devsite-nav-item"> <a href="https://www.youtube.com/@googlecloudtech" class="devsite-nav-title gc-analytics-event " data-category="Site-Wide Custom Events" data-label="Responsive Tab: Google Cloud Tech Youtube Channel" track-type="navMenu" track-metadata-eventDetail="globalMenu" track-metadata-position="nav"> <span class="devsite-nav-text" tooltip > Google Cloud Tech Youtube Channel </span> </a> </li> </ul> </div> </div> </nav> </devsite-book-nav> <section id="gc-wrapper"> <main role="main" id="main-content" class="devsite-main-content" has-book-nav has-sidebar > <div class="devsite-sidebar"> <div class="devsite-sidebar-content"> <devsite-toc class="devsite-nav" role="navigation" aria-label="On this page" depth="2" scrollbars ></devsite-toc> <devsite-recommendations-sidebar class="nocontent devsite-nav"> </devsite-recommendations-sidebar> </div> </div> <devsite-content> <article class="devsite-article"><style> /* Styles inlined from /docs/css/release-notes.css */ .release-announcement, .release-breaking, .release-changed, .release-deprecated, .release-feature, .release-fixed, .release-issue, .release-libraries, .release-security { margin-left: 110px; position: relative; } .release-announcement::before, .release-breaking::before, .release-changed::before, .release-deprecated::before, .release-feature::before, .release-fixed::before, .release-issue::before, .release-libraries::before, .release-security::before{ border-radius: 14px; display: inline-block; font-size: 12px; font-weight: 700; height: 28px; left: -110px; line-height: 28px; margin-right: 10px; position: absolute; text-align: center; text-transform: uppercase; top: -3px; width: 100px; } /* Colours for each type of changelog entry. */ .release-breaking::before { content: "breaking"; background-color: #db4437; color: #fff; } .release-changed::before { content: "change"; background-color: #f4b400; color: #fff; } .release-deprecated::before { content: "deprecated"; background-color: #db4437; color: #fff; } .release-feature::before { content: "feature"; background-color: #0f9d58; color: #fff; } .release-fixed::before { content: "fixed"; background-color: #4285f4; color: #fff; } .release-issue::before { content: "issue"; background-color: #db4437; color: #fff; } .release-security::before { content: "security"; background-color: #ff6e40; color: #fff; } .release-announcement::before { content: "announcement"; background-color: #ffbc00; color: #fff; font-size: 11px; } .release-libraries::before { content: "libraries"; background-color: #860ac6; color: #fff; } .release-note-sitemap-card { display: flex; flex-wrap: wrap; } .release-note-sitemap-card-item { flex: 1 50%; } </style> <div class="devsite-article-meta nocontent" role="navigation"> <ul class="devsite-breadcrumb-list" aria-label="Breadcrumb"> <li class="devsite-breadcrumb-item "> <a href="https://cloud.google.com/" class="devsite-breadcrumb-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Breadcrumbs" data-value="1" track-type="globalNav" track-name="breadcrumb" track-metadata-position="1" track-metadata-eventdetail="Google Cloud" > Home </a> </li> <li class="devsite-breadcrumb-item "> <div class="devsite-breadcrumb-guillemet material-icons" aria-hidden="true"></div> <a href="https://cloud.google.com/vertex-ai" class="devsite-breadcrumb-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Breadcrumbs" data-value="2" track-type="globalNav" track-name="breadcrumb" track-metadata-position="2" track-metadata-eventdetail="Vertex AI" > Vertex AI </a> </li> <li class="devsite-breadcrumb-item "> <div class="devsite-breadcrumb-guillemet material-icons" aria-hidden="true"></div> <a href="https://cloud.google.com/vertex-ai/docs" class="devsite-breadcrumb-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Breadcrumbs" data-value="3" track-type="globalNav" track-name="breadcrumb" track-metadata-position="3" track-metadata-eventdetail="Vertex AI" > Documentation </a> </li> <li class="devsite-breadcrumb-item "> <div class="devsite-breadcrumb-guillemet material-icons" aria-hidden="true"></div> <a href="https://cloud.google.com/vertex-ai/docs/release-notes" class="devsite-breadcrumb-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Breadcrumbs" data-value="4" track-type="globalNav" track-name="breadcrumb" track-metadata-position="4" track-metadata-eventdetail="" > Resources </a> </li> </ul> <devsite-thumb-rating position="header"> </devsite-thumb-rating> </div> <devsite-feedback position="header" project-name="Vertex AI" product-id="5164350" bucket="Documentation" context="" version="t-devsite-webserver-20250211-r00-rc00.466928320959134316" data-label="Send Feedback Button" track-type="feedback" track-name="sendFeedbackLink" track-metadata-position="header" class="nocontent" project-feedback-url="https://issuetracker.google.com/issues/new?component=1130925" project-icon="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/favicons/onecloud/super_cloud.png" project-support-url="https://cloud.google.com/support-hub" > <button> Send feedback </button> </devsite-feedback> <h1 class="devsite-page-title" tabindex="-1"> Vertex AI release notes </h1> <devsite-feature-tooltip ack-key="AckCollectionsBookmarkTooltipDismiss" analytics-category="Site-Wide Custom Events" analytics-action-show="Callout Profile displayed" analytics-action-close="Callout Profile dismissed" analytics-label="Create Collection Callout" class="devsite-page-bookmark-tooltip nocontent" dismiss-button="true" id="devsite-collections-dropdown" dismiss-button-text="Dismiss" close-button-text="Got it"> <devsite-bookmark></devsite-bookmark> <span slot="popout-heading"> Stay organized with collections </span> <span slot="popout-contents"> Save and categorize content based on your preferences. </span> </devsite-feature-tooltip> <div class="devsite-page-title-meta"><devsite-view-release-notes></devsite-view-release-notes></div> <devsite-toc class="devsite-nav" depth="2" devsite-toc-embedded > </devsite-toc> <div class="devsite-article-body clearfix "> <section class="intro"> <p>This page documents production updates to Vertex AI. You can periodically check this page for announcements about new or updated features, bug fixes, known issues, and deprecated functionality.</p> <p>See also:</p> <ul> <li><a href="/colab/docs/release-notes">Colab Enterprise release notes</a></li> <li><a href="/vertex-ai/generative-ai/docs/release-notes">Generative AI on Vertex AI release notes</a></li> <li><a href="/vertex-ai/docs/workbench/release-notes">Vertex AI Workbench release notes</a></li> </ul> </section> <section> <p> You can see the latest product updates for all of Google Cloud on the <a href="/release-notes" track-type="releaseNotes" track-name="releaseNotesIndex"> Google Cloud</a> page, browse and filter all release notes in the <a href="https://console.cloud.google.com/release-notes" target="console" class="external" track-type="releaseNotes" track-name="consoleLink">Google Cloud console</a>, or programmatically access release notes in <a class="external" href="https://console.cloud.google.com/bigquery?p=bigquery-public-data&d=google_cloud_release_notes&t=release_notes&page=table" track-type="releaseNotes" track-name="bigqueryDataset">BigQuery</a>. </p> </section> <section class="xml"> <p> To get the latest product updates delivered to you, add the URL of this page to your <a href="https://wikipedia.org/wiki/Comparison_of_feed_aggregators" class="external">feed reader</a>, or add the <a href="https://cloud.google.com/feeds/vertex-ai-product-group-release-notes.xml">feed URL</a> directly. </p> </section> <section class="releases"> <section class="releases"> <h2 id="February_12_2025" data-text="February 12, 2025" tabindex="-1">February 12, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="508e6542" class="release-feature"> <p>Deepseek-V3 and Deepseek-R1 have been added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a> in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>:</p> <ul> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/deepseek-ai/model-garden/deepseek-v3">DeepSeek-V3</a> (671B) is a powerful Mixture-of-Experts (MoE) language model with 671B total parameters with 37B activated for each token.</li> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/deepseek-ai/model-garden/deepseek-r1">DeepSeek-R1</a> (671B) is one of the first-generation reasoning models introduced by DeepSeek and offers performance comparable to OpenAI-o1 across math, code, and reasoning tasks.</li> </ul> <p>You can use a <a href="https://colab.research.google.com/github/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_deepseek_deployment.ipynb">notebook</a> to deploy these models.</p> </div> <h2 id="February_11_2025" data-text="February 11, 2025" tabindex="-1">February 11, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="731b867b" class="release-announcement"> <p>The <a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama-3.3-70b-instruct-maas">Llama 3.3 70B</a> model that is managed on Vertex AI is now in <a href="https://cloud.google.com/products?e=13802955&hl=en#product-launch-stages">Preview</a>.</p> </div> <h2 id="February_07_2025" data-text="February 07, 2025" tabindex="-1">February 07, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="667f413e" class="release-feature"> <p><a href="https://console.cloud.google.com/vertex-ai/publishers/deepseek-ai/model-garden/deepseek-r1">deepseek-ai/deepseek-r1</a> and <a href="https://console.cloud.google.com/vertex-ai/publishers/microsoft/model-garden/phi-4">microsoft/Phi-4</a> models were added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>.</p> </div> <div id="7f87fd1f" class="release-changed"> <p>The following advanced LLM inference optimization techniques are available in Model Garden in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>:</p> <ul> <li><p>Prefix caching reuses computations from previously generated text, eliminating redundant processing. It reduces time-to-first-token for requests with common prompt prefixes. Prefix caching is available for the following models:</p> <ul> <li>vLLM: Llama 3.1 (8b, 70b), Llama 3.3 (70b)</li> <li>Hex-LLM: Llama 2 (7b, 13b), Llama 3 (8b), Llama 3.1 (8b, 70b), Llama 3.2 (1b, 3b), Llama Guard (1b, 8b), CodeLlama (7b, 13b), Gemma (2b, 7b), CodeGemma (2b, 7b), Mistral-7B (v0.2, v0.3), Mixtral-8x7B (v0.1)</li> </ul></li> <li><p>Speculative decoding is an effective optimization technique to reduce generation time-per-output-token latency. For more information, see the <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_advanced_features.ipynb">Model Garden advanced features notebook</a>.</p></li> </ul> </div> <h2 id="February_05_2025" data-text="February 05, 2025" tabindex="-1">February 05, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="8923e49c" class="release-feature"> <p><strong>Gemini 2.0 Flash general availability for text-only output</strong></p> <p>Gemini 2.0 Flash is now generally available for text-only outputs. Multimodal outputs are still available only as a private preview. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/gemini-v2">Gemini 2.0</a>.</p> </div> <div id="d59f5bdc" class="release-feature"> <p><strong>New Gemini 2.0 Pro and Gemini 2.0 Flash-Lite models available to users</strong></p> <p>Two new models in the Gemini 2.0 family are now available to users:</p> <ul> <li><strong>Gemini 2.0 Pro:</strong> Our strongest model for coding and world knowledge, featuring a 2M long context window. Gemini 2.0 Pro is available as an experimental model in Vertex AI.</li> <li><strong>Gemini 2.0 Flash-Lite:</strong> Our fastest and most cost efficient Flash model. Gemini 2.0 Flash-Lite is available as a Preview model in Vertex AI.</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/gemini-v2">Gemini 2.0</a></p> </div> <h2 id="January_31_2025" data-text="January 31, 2025" tabindex="-1">January 31, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e45be167" class="release-feature"> <p>You can now monitor usage, throughput, and latency and troubleshoot 429 errors on Vertex AI foundation models, like Google Gemini and Anthropic Claude, by using a predefined dashboard. After querying a model from the <a href="https://cloud.google.com/model-garden">Vertex AI Model Garden</a>, you can find the name of the model you queried in the Vertex AI <strong>Dashboard</strong> page under the "Model observability" heading.</p> <p>To customize the dashboard and explore relevant metrics in Cloud Monitoring, click <strong>Show All Metrics</strong>. For information about using dashboards in Cloud Monitoring, see <a href="https://cloud.google.com/monitoring/charts/predefined-dashboards">View and customize Google Cloud dashboards</a>.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="8578817d" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/general/deployment#deploy_a_model_to_an_endpoint">Dedicated Public Endpoints </a>and <a href="https://cloud.google.com/vertex-ai/docs/predictions/private-service-connect">Private Service Connect Endpoints (PSC-E)</a> for Vertex AI Prediction are generally available (GA).</p> </div> <h2 id="January_30_2025" data-text="January 30, 2025" tabindex="-1">January 30, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9cd9d440" class="release-deprecated"> <p>Mistral Large (24.07) and Codestral (24.05) that are offered as a Model as a Service (MaaS) models in Model Garden are deprecated. For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/deprecations">Generative AI on Vertex AI deprecations</a>.</p> </div> <h2 id="January_29_2025" data-text="January 29, 2025" tabindex="-1">January 29, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="54adae48" class="release-feature"> <p><strong>New Imagen 3 image generation model available to users</strong></p> <p>A newer improved Imagen 3 image generation model is now available to all users: </p> <ul> <li><code translate="no" dir="ltr">imagen-3.0-generate-002</code></li> </ul> <p>This image generation model supports the following additional features:</p> <ul> <li><strong>Prompt enhancement</strong> - The LLM-based prompt rewriter tool adds additional details and descriptive language to the prompt you provide, generally resulting in higher quality generated images. This feature is configurable and is enabled by default.</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/model-versioning">Imagen on Vertex AI model versions and lifecycle</a> and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/generate-images">Generate images using text prompts</a>. </p> </div> <h2 id="January_22_2025" data-text="January 22, 2025" tabindex="-1">January 22, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="98b97531" class="release-announcement"> <p><strong>LangChain on Vertex AI</strong></p> <p>Billing for LangChain on Vertex AI will start on March 4, 2025. </p> <p>The pricing structure is based on vCPU hours and GiB hours used. This means that you will be charged for both the compute (vCPU) and memory resources consumed by your LangChain on Vertex AI workloads.</p> <p>You can review the pricing details in the table below.</p> <table> <thead> <tr> <th>Product</th> <th>SKU ID</th> <th>Price</th> </tr> </thead> <tbody> <tr> <td>ReasoningEngine vCPU</td> <td>8A55-0B95-B7DC</td> <td>$0.0994/vCPU-Hr</td> </tr> <tr> <td>ReasoningEngine Memory</td> <td>0B45-6103-6EC1</td> <td>$0.0105/GiB-Hr</td> </tr> </tbody> </table> </div> <h2 id="January_21_2025" data-text="January 21, 2025" tabindex="-1">January 21, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="681363fd" class="release-deprecated"> <p>Anthropic's Claude 3 Sonnet that is offered as a Model as a Service (MaaS) model in Model Garden is deprecated. For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/deprecations">Generative AI on Vertex AI deprecations</a>.</p> </div> <h2 id="January_17_2025" data-text="January 17, 2025" tabindex="-1">January 17, 2025</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e73e0bdd" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/evaluation-agents">Agent evaluation</a> using the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/evaluation-overview">Gen AI evaluation service</a> is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <h2 id="January_16_2025" data-text="January 16, 2025" tabindex="-1">January 16, 2025</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="865217b0" class="release-feature"> <p><strong>M127 release</strong></p> <p>The M127 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Fixed an issue related to ownership of the home directory when using authorized ssh keys.</li> </ul> </div> <div id="9782c84c" class="release-feature"> <p>The M127 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed an issue related to ownership of the home directory when using authorized ssh keys.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="d88b197d" class="release-feature"> <p><strong>M127 release</strong></p> <p>The M127 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Fixed an issue related to ownership of the home directory when using authorized ssh keys.</li> </ul> </div> <h2 id="December_20_2024" data-text="December 20, 2024" tabindex="-1">December 20, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a785ff54" class="release-changed"> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/rag-overview">RAG Engine</a> is <a href="https://cloud.google.com/products#product-launch-stages">generally available (GA)</a>.</p> <p>The supported models include the following:</p> <ul> <li>Google Gemini</li> <li>Google embedding and OSS E5 embedding models</li> <li>Model Garden self-deployed OSS LLMs</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-partner-models">Model as a service (MaaS) Llama models</a></li> </ul> <p>The supported features include the following:</p> <ul> <li><strong><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/use-data-connectors">Data connectors</a></strong>: Google Cloud Storage, Google Drive, Slack, Jira, and SharePoint</li> <li><strong><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/supported-documents">Document types</a></strong>: Google Workspace documents, HTML, JSON, Markdown, PDF, and text files</li> <li><strong>Transformations</strong>: fixed-size chunking and chunk overlap</li> <li><strong>Vector databases</strong>: Vertex AI Vector Search and Pinecone</li> </ul> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f0c08052" class="release-feature"> <p><strong>Vector Search hybrid search and sparse embeddings are generally available (GA)</strong></p> <p>Vector Search hybrid search and sparse embeddings are generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). Hybrid search uses both dense and sparse embeddings, which lets you search based on a combination of keyword search and semantic search. For more information about hybrid search, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/about-hybrid-search">About hybrid search</a>.</p> </div> <h2 id="December_18_2024" data-text="December 18, 2024" tabindex="-1">December 18, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="3cefa1a2" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-hex-llm">Hex-LLM: High-Efficiency Large Language Model Serving</a> is available in <a href="https://cloud.google.com/products#product-launch-stages">General Availability (GA)</a>.</p> <p>This launch adds support for the following models: </p> <ul> <li>Llama 3.1 </li> <li>Llama 3.2 </li> <li>Phi-3 </li> <li>Qwen2 and Qwen2.5 </li> </ul> <p>Additional supported features:</p> <ul> <li>Multi-host serving. </li> <li>Disaggregated serving (experimental). </li> <li>Prefix caching. </li> <li>AWQ quantization.</li> </ul> </div> <h2 id="December_17_2024" data-text="December 17, 2024" tabindex="-1">December 17, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="681bddd6" class="release-feature"> <p>You can copy tuned Gemini 1.5 Pro 002 and Gemini 1.5 Flash 002 adapter models across projects. For details, see <a href="https://cloud.google.com/vertex-ai/docs/model-registry/copy-model">Copy a model in Vertex AI Model Registry</a>.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="cadc0dcd" class="release-feature"> <p>You can copy tuned Gemini 1.5 Pro 002 and Gemini 1.5 Flash 002 adapter models across projects. For details, see <a href="https://cloud.google.com/vertex-ai/docs/model-registry/copy-model">Copy a model in Vertex AI Model Registry</a>.</p> </div> <h2 id="December_11_2024" data-text="December 11, 2024" tabindex="-1">December 11, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f9df6930" class="release-feature"> <p>The Gemini 2.0 Flash (<code translate="no" dir="ltr">gemini-2.0-flash-exp</code>) model is Generally available for grounded answer generation with RAG. This model is tuned to address context-based question and answering tasks. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/ground-gemini">Ground responses for Gemini models</a>.</p> </div> <h2 id="December_10_2024" data-text="December 10, 2024" tabindex="-1">December 10, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="483d7fbb" class="release-feature"> <p>Gemini in Colab Enterprise, which is a product in the Gemini for Google Cloud portfolio, now includes error fixing in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. Gemini in Colab Enterprise can suggest fixes when your code produces errors. For more information, see <a href="https://cloud.google.com/colab/docs/use-code-completion#fix-errors">Fix errors</a>.</p> <p>To enable and activate Gemini in Colab Enterprise features, see <a href="https://cloud.google.com/gemini/docs/colab/set-up-gemini">Set up Gemini in Colab Enterprise</a>.</p> </div> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="884d573b" class="release-feature"> <p><strong>Imagen 3 image generation models Generally Available to all users</strong></p> <p>Imagen 3 image generation models are now available to all users without requiring prior approval. These include the following image generation models:</p> <ul> <li><code translate="no" dir="ltr">imagen-3.0-generate-001</code></li> <li><code translate="no" dir="ltr">imagen-3.0-fast-generate-001</code> (low latency model)</li> </ul> <p>Prior image generation models (<code translate="no" dir="ltr">imagegeneration@006</code>, <code translate="no" dir="ltr">imagegeneration@005</code>, <code translate="no" dir="ltr">imagegeneration@002</code>) still require approval to use.</p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/model-versioning">Imagen on Vertex AI model versions and lifecycle</a> and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/generate-images">Generate images using text prompts</a>.</p> </div> <div id="1faac5b8" class="release-feature"> <p><strong>Imagen 3 Customization model Generally Available to approved users</strong></p> <p>Imagen 3 Customization model is now available to approved users. This includes the following model:</p> <ul> <li><code translate="no" dir="ltr">imagen-3.0-capability</code></li> </ul> <p>Imagen 3 Customization lets you guide image generation by providing reference images (few-shot learning). Imagen 3 Customization lets you customize generated images for the following feature categories:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/subject-customization">Subject Customization</a> (product, person, and animal companion)</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/style-customization">Style Customization</a></li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-controlled">Controlled Customization</a> (canny edge and scribble)</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/instruct-customization">Instruct Customization</a> (Style transfer)</li> </ul> </div> <div id="c3ec4d4e" class="release-feature"> <p><strong>Imagen 3 editing model Generally Available to approved users</strong></p> <p>The Imagen 3 Editing model is now available to approved users. This includes the following model:</p> <ul> <li><code translate="no" dir="ltr">imagen-3.0-capability</code></li> </ul> <p>This model offers the following additional features:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-inpainting">Inpainting</a> - Add or remove content from a masked area of an image</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-outpainting">Outpainting</a> - Expand a masked area of an image</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-product-image">Product image editing</a> - Identify and maintain a primary product while changing the background or product position</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/model-versioning">Model versions</a>.</p> </div> <h2 id="December_06_2024" data-text="December 06, 2024" tabindex="-1">December 06, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="74ab9f7d" class="release-security"> <p>A vulnerability was discovered in the Vertex AI API serving Gemini multimodal requests, allowing bypass of VPC Service Controls. For details, see the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/security-bulletins#gcp-2024-063">Security bulletins</a> page.</p> </div> <h2 id="December_02_2024" data-text="December 02, 2024" tabindex="-1">December 02, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="af2308f3" class="release-feature"> <p><strong><a href="https://cloud.google.com/products/#product-launch-stages">Preview</a></strong>: You can consume reservations of VMs that have GPUs attached with your custom training jobs or prediction jobs. Reservations of Compute Engine zonal resources help you gain a high level of assurance that your jobs have the necessary resources to run. For more information, see the following:</p> <ul> <li><p><a href="https://cloud.google.com/vertex-ai/docs/training/use-reservations">Use reservations with training</a></p></li> <li><p><a href="https://cloud.google.com/vertex-ai/docs/predictions/use-reservations">Use reservations with prediction</a></p></li> </ul> </div> <h2 id="November_21_2024" data-text="November 21, 2024" tabindex="-1">November 21, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="54e8b80c" class="release-announcement"> <p>Mistral Large (24.11) is Generally Available on Vertex AI as a managed model. To learn more, view the <a href="https://console.cloud.google.com/vertex-ai/publishers/mistralai/model-garden/mistral-large-2411">Mistral Large (24.11) model card</a> in Model Garden.</p> </div> <div id="14dda96c" class="release-feature"> <p>The Gen AI evaluation service can now help you evaluate your translation models using MetricX, COMET, and BLEU metrics. To learn more about evaluating your translation models, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/determine-eval#translation-models">Evaluate translation models</a>.</p> </div> <h2 id="November_20_2024" data-text="November 20, 2024" tabindex="-1">November 20, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="0338e086" class="release-feature"> <p><strong>M126 release</strong> </p> <p>The M126 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Upgraded JupyterLab to 3.6.8.</li> <li>One or more framework versions have reached their end of patch and support dates. To view end of patch and support dates, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/images#supported-frameworks">Supported framework versions</a>. To create an instance using an image family that has reached its end of patch and support date, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/create-after-end-of-support">Create an instance after end of patch and support date</a>.</li> </ul> </div> <div id="bdc21ba0" class="release-feature"> <p>The M126 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Upgraded JupyterLab to 3.6.8.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="5200c22d" class="release-feature"> <p><strong>M126 release</strong></p> <p>The M126 release of Vertex AI Workbench instances includes the following:</p> <ul> <li><a href="https://cloud.google.com/products#product-launch-stages">Preview</a>: JupyterLab 4+ is available on new Vertex AI Workbench instances. To try it, select JupyterLab 4 when you <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/create#create">create your instance</a>.</li> <li>Upgraded JupyterLab to 3.6.8.</li> </ul> </div> <h2 id="November_08_2024" data-text="November 08, 2024" tabindex="-1">November 08, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1561d713" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/llama-batch">Batch predictions for Llama models</a> on Vertex AI (MaaS) is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <div id="721402c1" class="release-feature"> <p><strong>Batch prediction support for Gemini</strong></p> <p>Batch prediction is available for Gemini in <a href="https://cloud.google.com/products#product-launch-stages">General Availability (GA)</a>. Available Gemini models include Gemini 1.0 Pro, Gemini 1.5 Pro, and Gemini 1.5 Flash. To get started with batch prediction, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/batch-prediction-gemini">Get batch predictions for Gemini</a>.</p> </div> <h2 id="November_05_2024" data-text="November 05, 2024" tabindex="-1">November 05, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b6f3959a" class="release-changed"> <p>We are extending the availability of Gemini 1.0 Pro 001 and Gemini 1.0 Pro Vision 001 from February 15, 2025 to April 9, 2025. For details, see the <a href="https://cloud.devsite.corp.google.com/vertex-ai/generative-ai/docs/deprecations">Deprecations</a>.</p> </div> <h2 id="November_04_2024" data-text="November 04, 2024" tabindex="-1">November 04, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="878001af" class="release-changed"> <p>The translation LLM now supports Polish, Turkish, Indonesian, Dutch, Vietnamese, Thai and Czech. For the full list of supported languages, see the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/translate/translate-text">Translate text</a> page.</p> </div> <div id="acc903ef" class="release-announcement"> <p>The Anthropic Claude Haiku 3.5 is Generally Available on Vertex AI. To learn more, view the <a href="https://console.cloud.google.com/vertex-ai/publishers/anthropic/model-garden/claude-3-5-haiku">Claude Haiku 3.5 model card</a> in Model Garden.</p> </div> <h2 id="October_31_2024" data-text="October 31, 2024" tabindex="-1">October 31, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a884f10c" class="release-feature"> <p>PSC-I Egress is supported for Ray clusters Vertex AI. PSC-I is recommended for private connectivity since it reduces the chance of IP exhaustion, and allows for transitive peering. Check out <a href="https://cloud.google.com/vertex-ai/docs/open-source/ray-on-vertex-ai/create-cluster#use-psc-i-egress">Private Service Connect interface for Ray on Vertex AI</a>. This feature is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <div id="aa0f9463" class="release-feature"> <p>Private Service Connect interface (PSC-I) is now supported for ML pipeline runs in Vertex AI Pipelines. PSC-I is recommended for private connectivity, since it reduces the chance of IP exhaustion, and allows for transitive peering.</p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/configure-private-service-connect">Configure Private Service Connect interface for a pipeline</a>. This feature is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <h2 id="October_28_2024" data-text="October 28, 2024" tabindex="-1">October 28, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f40c76d0" class="release-feature"> <p>You can now fine-tune the following models from the Cloud console:</p> <ul> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">Gemma 2</a></li> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama3_1">Llama 3.1</a></li> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/mistral-ai/model-garden/mistral">Mistral</a></li> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/mistral-ai/model-garden/mixtral">Mixtral</a></li> </ul> </div> <div id="1d8a39a8" class="release-feature"> <p>The <a href="https://console.cloud.google.com/vertex-ai/publishers/openai/model-garden/whisper-large">Whisper large v3 and Whisper large v3 turbo</a> models have been added to Model Garden.</p> </div> <div id="f91150ef" class="release-changed"> <p>Updated the fine-tuning notebooks for <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">Gemma 2</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama3_1">Llama 3.1</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/mistral-ai/model-garden/mistral">Mistral</a>, and <a href="https://console.cloud.google.com/vertex-ai/publishers/mistral-ai/model-garden/mixtral">Mixtral</a> with the following enhancements:</p> <ul> <li>The notebooks use an updated high-performance container for single host multi-GPU LoRA fine-tuning. <ul> <li>Better throughput and GPU utilization with well-tested max-sequence-lengths.</li> <li>Support for input token masking.</li> <li>No out of memory (OOM) error during fine-tuning.</li> </ul></li> <li>Added a custom dataset example that uses a template and format validation.</li> <li>Support for a default accelerator pool with quota checks.</li> <li>Improved documentation.</li> </ul> </div> <h2 id="October_25_2024" data-text="October 25, 2024" tabindex="-1">October 25, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="72e8b9bf" class="release-feature"> <p>Colab Enterprise is now available in the following regions:</p> <ul> <li>Hamina, Finland (<code translate="no" dir="ltr">europe-north1</code>)</li> <li>Milan, Italy (<code translate="no" dir="ltr">europe-west8</code>)</li> <li>Tel Aviv, Israel (<code translate="no" dir="ltr">me-west1</code>)</li> <li>Warsaw, Poland (<code translate="no" dir="ltr">europe-central2</code>)</li> </ul> <p>See <a href="https://cloud.google.com/colab/docs/locations">Colab Enterprise locations</a>.</p> </div> <h2 id="October_22_2024" data-text="October 22, 2024" tabindex="-1">October 22, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bced24f0" class="release-announcement"> <p>The Anthropic Claude Sonnet 3.5 v2 is Generally Available. To learn more, view the <a href="https://console.cloud.google.com/vertex-ai/publishers/anthropic/model-garden/claude-3-5-sonnet-v2">Claude Sonnet 3.5 v2 model card</a> in Model Garden.</p> </div> <h2 id="October_18_2024" data-text="October 18, 2024" tabindex="-1">October 18, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a3b8dbb7" class="release-announcement"> <p>The <a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama-3.1-405b-instruct-maas">Llama 3.1 405B</a> model that is managed on Vertex AI is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>.</p> </div> <h2 id="October_09_2024" data-text="October 09, 2024" tabindex="-1">October 09, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="eec8bb75" class="release-feature"> <p>The Vertex AI Gemini API SDK supports tokenization capabilities for local token counting and computation. This is a streamlined way to compute tokens locally, ensuring compatibility across different Gemini models and their tokenizers. Supported models include <code translate="no" dir="ltr"> gemini-1.5-flash</code> and <code translate="no" dir="ltr"> gemini-1.5-pro</code> . To learn more, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/list-token">Count tokens</a>.</p> </div> <h2 id="October_08_2024" data-text="October 08, 2024" tabindex="-1">October 08, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1a5d7138" class="release-feature"> <p><strong>Vector Search Private Service Connect automation</strong></p> <p>Deploying an index with Private Service Connect automation is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). You can set up a service connection policy so that you don't have to manually create a compute address and forwarding rule after each index deployment. </p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/setup/private-service-connect">Set up Vector Search with Private Service Connect</a>.</p> </div> <h2 id="October_04_2024" data-text="October 04, 2024" tabindex="-1">October 04, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="293808a7" class="release-feature"> <p>The AI assistant in Vertex AI Studio can help you refine and generate prompts. This feature is in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. To learn more, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/prompts/ai-powered-prompt-writing">Use AI-powered prompt writing tools</a>.</p> </div> <div id="736463cc" class="release-feature"> <p><a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/prompt-guard">Prompt Guard</a> and <a href="https://console.cloud.google.com/vertex-ai/publishers/black-forest-labs/model-garden/flux1-schnell">Flux</a> were added to <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/model-garden/explore-models">Model Garden</a>.</p> </div> <div id="99807f91" class="release-feature"> <p>You can deploy Hugging Face models on Google Cloud that have <a href="https://huggingface.github.io/text-embeddings-inference/">text embedding inference</a> enabled or <a href="https://huggingface.co/docs/inference-endpoints/supported_tasks">pytorch inference</a> enabled. For more information, see the <a href="http://console.cloud.google.com/vertex-ai/model-garden;action=deploy;hfSource=true">Hugging Face model deployment</a> in the console.</p> </div> <div id="bcc48b60" class="release-changed"> <p>Added multiple deployment settings (with A100-80G and H100) and sample requests for some popular models, including <a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama3_1">Llama 3.1</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">Gemma 2</a>, and <a href="https://console.cloud.google.com/vertex-ai/publishers/mistral-ai/model-garden/mixtral">Mixtral</a>. </p> </div> <div id="86a249e3" class="release-changed"> <p>Added dynamic LoRA serving for <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_llama3_1_deployment.ipynb">Llama 3.1</a> and <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_xl_lora.ipynb">Stable Diffusion XL</a>.</p> </div> <h2 id="October_03_2024" data-text="October 03, 2024" tabindex="-1">October 03, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="fa765709" class="release-feature"> <p>Gemini in Colab Enterprise, which is a product in the Gemini for Google Cloud portfolio, is <a href="https://cloud.google.com/products/#product-launch-stages">generally available</a>. Gemini in Colab Enterprise helps you write code by suggesting code as you type. You can also use the Help me code tool to generate code from a description of what you want.</p> <p>Gemini in Colab Enterprise is available to try at no cost through December 31, 2024.</p> <p>To learn how to enable and activate Gemini in Colab Enterprise features, see <a href="https://cloud.google.com/gemini/docs/colab/set-up-gemini">Set up Gemini in Colab Enterprise</a>.</p> </div> <h2 id="October_01_2024" data-text="October 01, 2024" tabindex="-1">October 01, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fc0feb06" class="release-feature"> <p><strong>Grounding: Dynamic retrieval for grounded results (GA)</strong></p> <p>Dynamic retrieval lets you choose when to turn off grounding with Google Search. This is useful when a prompt doesn't require an answer grounded in Google Search, and the supported models can provide an answer based on their knowledge without grounding. Dynamic retrieval helps you manage latency, quality, and cost more effectively.</p> <p>This feature is <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/ground-gemini#dynamic-retrieval">Dynamic retrieval</a>.</p> </div> <h2 id="September_30_2024" data-text="September 30, 2024" tabindex="-1">September 30, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6374d3bd" class="release-feature"> <p>Prompt templates let you to test how different prompt formats perform with different sets of prompt data. This feature is in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. To learn more, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/prompts/prompt-templates">Use prompt templates</a>.</p> </div> <h2 id="September_26_2024" data-text="September 26, 2024" tabindex="-1">September 26, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="b59d9c95" class="release-feature"> <p><strong>M125 release</strong> </p> <p>The M125 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Patched a vulnerability with <code translate="no" dir="ltr">adm</code> and <code translate="no" dir="ltr">docker</code> permissions when the instance's root access isn't enabled.</li> </ul> </div> <div id="bb4d8fb8" class="release-feature"> <p>The M125 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Patched a vulnerability with <code translate="no" dir="ltr">adm</code> and <code translate="no" dir="ltr">docker</code> permissions when the instance's root access isn't enabled.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="d0c89f2b" class="release-feature"> <p><strong>M125 release</strong> </p> <p>The M125 release of Vertex AI Workbench instances includes the following:</p> <ul> <li><code translate="no" dir="ltr">bigframes</code> 1.9.0 is now available in all environments except TensorFlow.</li> <li>Fixed a regression introduced in M124 where Conda was getting downgraded to an older version.</li> <li>Patched a vulnerability with <code translate="no" dir="ltr">adm</code> and <code translate="no" dir="ltr">docker</code> permissions when the instance's root access isn't enabled.</li> </ul> </div> <h2 id="September_25_2024" data-text="September 25, 2024" tabindex="-1">September 25, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9783b493" class="release-announcement"> <p>The Llama 3.2 90B model is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a> on Vertex AI. Llama 3.2 90B enables developers to build and deploy the latest generative AI models and applications that use Llama's capabilities, such as image reasoning. Llama 3.2 is also designed to be more accessible for on-device applications. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/llama">Llama models</a>. </p> </div> <h2 id="September_24_2024" data-text="September 24, 2024" tabindex="-1">September 24, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="25df022e" class="release-announcement"> <p>New stable versions of Gemini 1.5 Pro (<code translate="no" dir="ltr">gemini-1.5-pro-002</code>) and Gemini 1.5 Flash (<code translate="no" dir="ltr">gemini-1.5-flash-002</code>) are <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>. These models introduce broad quality improvements over the previous <code translate="no" dir="ltr">001</code> versions, with significant gains in the following categories:</p> <ul> <li>Factuality and reduce model hallucinations</li> <li>Openbook Q&A for RAG use cases</li> <li>Instruction following</li> <li>Multilingual understanding in 102 languages, especially in Korean, French, German, Spanish, Japanese, Russian, and Chinese.</li> <li>SQL generation</li> <li>Audio understanding</li> <li>Document understanding</li> <li>Long context</li> <li>Math and reasoning</li> </ul> <p>For more information about differences with the previous model versions, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versions">Model versions and lifecycle</a>.</p> </div> <div id="d2d3256e" class="release-announcement"> <p>The 2M context window with Gemini 1.5 Pro is now in <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>, which opens up long-form multimodal use cases that only Gemini can support.</p> </div> <div id="0ba9805a" class="release-feature"> <p>Use Gemini to directly analyze YouTube videos and publicly available media (such as images, audio, and video) by using a link. This feature is in <a href="https://cloud.google.com/products#product-launch-stages">Public Preview</a>.</p> </div> <div id="6a27c7e8" class="release-feature"> <p>The new API parameters <code translate="no" dir="ltr">audioTimestamp</code>, <code translate="no" dir="ltr">responseLogprob</code>, and <code translate="no" dir="ltr">logprobs</code> are in <a href="https://cloud.google.com/products#product-launch-stages">Public Preview</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/inference">API reference</a>.</p> </div> <div id="c58be0cc" class="release-feature"> <p>Gemini 1.5 Pro and Gemini 1.5 Flash now support multimodal input with <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/function-calling">function calling</a>. This feature is in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <div id="68f405ae" class="release-feature"> <p>The Vertex AI prompt optimizer adapts your prompts using the optimal instructions and examples to elicit the best performance from your chosen model. This feature is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. To learn more, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/prompts/prompt-optimizer">Optimize prompts</a>.</p> </div> <div id="5281c789" class="release-changed"> <p>Gemini 1.5 Pro and Gemini 1.5 Flash Tuning is now available in <a href="https://cloud.google.com/products#product-launch-stages">GA</a>. Tune Gemini with text, image, audio, and document data types using the latest models:</p> <ul> <li><code translate="no" dir="ltr">gemini-1.5-pro-002</code></li> <li><code translate="no" dir="ltr">gemini-1.5-flash-002</code></li> </ul> <p>Gemini 1.0 tuning remains in preview.</p> <p>For more information on tuning Gemini, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/gemini-use-supervised-tuning">Tune Gemini models by using supervised fine-tuning</a>.</p> </div> <div id="506cb0d8" class="release-changed"> <p>The latest versions of <strong>Gemini 1.5 Flash</strong> (<code translate="no" dir="ltr">gemini-1.5-flash-002</code>) and <strong>Gemini 1.5 Pro</strong> (<code translate="no" dir="ltr">gemini-1.5-pro-002</code>) use <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/dsq">dynamic shared quota</a>, which distributes on-demand capacity among all queries being processed. Dynamic shared quota is <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>.</p> </div> <div id="34d40fa7" class="release-announcement"> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/control-generated-output">Controlled generation</a> is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>.</p> </div> <h2 id="September_23_2024" data-text="September 23, 2024" tabindex="-1">September 23, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="aad21423" class="release-feature"> <p>You can now use customer-managed encryption keys (CMEK) to protect notebooks in Colab Enterprise.</p> <p>For more information, see <a href="https://cloud.google.com/colab/docs/cmek">Use customer-managed encryption keys</a>.</p> </div> <h2 id="September_20_2024" data-text="September 20, 2024" tabindex="-1">September 20, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f3f14b99" class="release-feature"> <p>Add label metadata to <code translate="no" dir="ltr">generateContent</code> and <code translate="no" dir="ltr">streamGenerateContent</code> API calls. For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/add-labels-to-api-calls">Add labels to API calls</a>.</p> </div> <h2 id="September_18_2024" data-text="September 18, 2024" tabindex="-1">September 18, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6c9860f7" class="release-announcement"> <p>Model Garden supports an organization policy so that administrators can limit access to certain models and capabilities. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/control-model-access">Control access to Model Garden models</a></p> </div> <h2 id="September_17_2024" data-text="September 17, 2024" tabindex="-1">September 17, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4cdf753d" class="release-feature"> <p>To ensure that VM resources are available when your custom training and prediction jobs need them, you can now use Compute Engine reservations. Reservations provide a high level of assurance in obtaining capacity for Compute Engine resources. This feature is available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a> for A2 and A3 machine series reservations.</p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/training/use-reservations">Use reservations with training</a> and <a href="https://cloud.google.com/vertex-ai/docs/predictions/use-reservations">Use reservations with prediction</a>.</p> </div> <div id="696518b1" class="release-feature"> <p>To reduce the cost of running your training and prediction jobs, you can now use Spot VMs. Spot VMs are virtual machine (VM) instances that are excess Compute Engine capacity. Spot VMs have significant discounts, but Compute Engine might preemptively stop or delete Spot VMs to reclaim the capacity at any time. This feature is available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>.</p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/training/use-spot-vms">Use Spot VMs with training</a> and <a href="https://cloud.google.com/vertex-ai/docs/predictions/use-spot-vms">Use Spot VMs with prediction</a>.</p> </div> <h2 id="September_16_2024" data-text="September 16, 2024" tabindex="-1">September 16, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b430cad1" class="release-feature"> <p>Schedule Vertex AI custom training jobs based on resource availability. For details, see the <a href="https://cloud.google.com/vertex-ai/docs/training/schedule-jobs-dws">Vertex AI documentation</a>.</p> </div> <h2 id="September_10_2024" data-text="September 10, 2024" tabindex="-1">September 10, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="f210f691" class="release-feature"> <p>The ability to back up and restore data on a Vertex AI Workbench instance is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/restore">Back up and restore an instance</a>.</p> </div> <h2 id="September_09_2024" data-text="September 09, 2024" tabindex="-1">September 09, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="60edaf13" class="release-feature"> <p>Ray cluster's autoscaling feature is now supported. See <a href="https://cloud.google.com/vertex-ai/docs/open-source/ray-on-vertex-ai/scale-clusters#autoscaling">Scale Ray clusters on Vertex AI</a></p> </div> <h2 id="September_03_2024" data-text="September 03, 2024" tabindex="-1">September 03, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c79610a3" class="release-changed"> <p>Gemini 1.5 Flash (<code translate="no" dir="ltr">gemini-1.5-flash</code>) supports <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/control-generated-output">controlled generation</a>.</p> </div> <h2 id="August_30_2024" data-text="August 30, 2024" tabindex="-1">August 30, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1a94c744" class="release-feature"> <p>Gen AI Evaluation Service is Generally Available. To learn more, see the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/evaluation-overview">Gen AI Evaluation Service overview</a>.</p> </div> <h2 id="August_26_2024" data-text="August 26, 2024" tabindex="-1">August 26, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="3c69c3b6" class="release-changed"> <p>For controlled generation, you can have the model respond with an enum value in plain text, as defined in your response schema. Set the <code translate="no" dir="ltr">responseMimeType</code> to <code translate="no" dir="ltr">text/x.enum</code>. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/control-generated-output">Control generated output</a>.</p> </div> <h2 id="August_22_2024" data-text="August 22, 2024" tabindex="-1">August 22, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a92e9965" class="release-changed"> <p><strong>AI21 Labs</strong></p> <p>Managed models from AI21 Labs are available on Vertex AI. To use a AI21 Labs model on Vertex AI, send a request directly to the Vertex AI API endpoint. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/ai21">AI21 models</a>.</p> </div> <h2 id="August_20_2024" data-text="August 20, 2024" tabindex="-1">August 20, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="63b245c2" class="release-feature"> <p><strong>M124 release</strong> </p> <p>The M124 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Pytorch 2.3.0 with CUDA 12.1 and Python 3.10 user-managed notebooks instances are now available.</li> <li>Fixed a bug that prevented kernels from appearing when the Cloud Resource Manager API is turned off and Dataproc is enabled.</li> </ul> </div> <h2 id="August_19_2024" data-text="August 19, 2024" tabindex="-1">August 19, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="201ddc8c" class="release-announcement"> <p>The ability to create a Vertex AI Workbench instance based on a custom container is now <a href="https://cloud.google.com/products#product-launch-stages">generally available</a>. Only custom containers derived from the Google-provided base container are supported. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/create-custom-container">Create an instance using a custom container</a>.</p> </div> <h2 id="August_14_2024" data-text="August 14, 2024" tabindex="-1">August 14, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="bce09aae" class="release-feature"> <p>The notebook scheduler is now <a href="https://cloud.google.com/products/#product-launch-stages">generally available</a>. See <a href="https://cloud.google.com/colab/docs/schedule-notebook-run">Schedule a notebook run</a>.</p> </div> <h2 id="August_11_2024" data-text="August 11, 2024" tabindex="-1">August 11, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c60f3d0a" class="release-changed"> <p>Generative AI on Vertex AI supports CMEK, VPC Service Controls, and Data Residency. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/general/features#security_controls_support_on">Security controls</a>.</p> </div> <h2 id="August_09_2024" data-text="August 09, 2024" tabindex="-1">August 09, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1848e997" class="release-announcement"> <p>Gemini on Vertex AI supports multiple response candidates. For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/inference">Generate content with the Gemini API</a>.</p> </div> <h2 id="August_08_2024" data-text="August 08, 2024" tabindex="-1">August 08, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="274cd5b9" class="release-feature"> <p><strong>M124 release</strong> </p> <p>The M124 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a bug that prevented kernels from appearing when the Cloud Resource Manager API is turned off and Dataproc is enabled.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="af425e98" class="release-feature"> <p><strong>M124 release</strong></p> <p>The M124 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Fixed a bug that prevented kernels from appearing when the Cloud Resource Manager API is turned off and Dataproc is enabled.</li> <li>Spark notebooks on Dataproc: The Serverless Spark runtime template creation screen now has an easy-to-use UI for configuring resource allocation, autoscaling, and GPU settings.</li> </ul> </div> <h2 id="August_05_2024" data-text="August 05, 2024" tabindex="-1">August 05, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="943c3aef" class="release-fixed"> <p>Fixed an issue in which users weren't able to access the Colab Enterprise UI when Colab Service Status was <strong>OFF for everyone</strong> in Google Workspace.</p> </div> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="24bea369" class="release-changed"> <p>The translation LLM now supports Arabic, Hindi, and Russian. For the full list of supported languages, see the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/translate/translate-text">Translate text</a> page.</p> </div> <h2 id="August_02_2024" data-text="August 02, 2024" tabindex="-1">August 02, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="8502a578" class="release-feature"> <p>Vertex AI SDK for Python supports token listing and counting for prompts without the need to make API calls. This feature is available in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/list-token">List and count tokens</a>.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4bf7bfa7" class="release-feature"> <p>The Vertex AI Model Registry now offers <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a> support for model copy across different projects. For information about how to copy your model projects and regions, see <a href="https://cloud.google.com/vertex-ai/docs/model-registry/copy-model">Copy models in Model Registry</a>.</p> </div> <h2 id="July_31_2024" data-text="July 31, 2024" tabindex="-1">July 31, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c77755f8" class="release-feature"> <p><strong>New Imagen on Vertex AI image generation model and features</strong></p> <p>The Imagen 3 image generation models (<code translate="no" dir="ltr">imagen-3.0-generate-001</code> and the low-latency version <code translate="no" dir="ltr">imagen-3.0-fast-generate-001</code>) are Generally Available to <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/overview#feature-launch-stage">approved users</a>. These models offer the following additional features:</p> <ul> <li>Additional aspect ratios (1:1, 3:4, 4:3, 9:16, 16:9)</li> <li>Digital watermark (<a href="https://deepmind.google/technologies/synthid/">SynthID</a>) enabled by default</li> <li>Watermark verification</li> <li>User-configurable safety features (safety setting, person/face setting)</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/model-versioning">Model versions</a> and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/generate-images">Generate images using text prompts</a>.</p> </div> <div id="1a1273bf" class="release-announcement"> <p>Gemma 2 2B is available in Model Garden. For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-gemma">Use Gemma open models</a>.</p> </div> <div id="e0fca3cb" class="release-feature"> <p>The following models have been added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>:</p> <ul> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">Gemma 2 2B</a>: A foundation LLM by Google Deepmind.</li> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/qwen/model-garden/qwen2">Qwen2</a>: An LLM series by Alibaba Cloud.</li> <li><a href="https://console.cloud.google.com/vertex-ai/publishers/microsoft/model-garden/phi3">Phi-3</a>: An LLM series by Microsoft.</li> </ul> </div> <div id="366b5a9d" class="release-changed"> <p>Resource and deployment settings were made to the following models:</p> <ul> <li>Added GPU inferences for <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">gemma2-27b</a> and <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">gemma2-27b-it</a> with verified performances.</li> <li>Added verified deployment settings for Mistral AI models that are deployed from Huggingface, including <a href="https://console.cloud.google.com/vertex-ai/publishers/mistralai/model-garden/mistral-nemo-instruct-2407;action=deploy;hfSource=true">mistralai/mistral-nemo-instruct-2407</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/mistralai/model-garden/mistral-nemo-base-2407;action=deploy;hfSource=true">mistralai/mistral-nemo-base-2407</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/mistralai/model-garden/mistral-large-instruct-2407;action=deploy;hfSource=true">mistralai/mistral-large-instruct-2407</a>, and <a href="https://console.cloud.google.com/vertex-ai/publishers/mistralai/model-garden/codestral-22b-v0.1;action=deploy;hfSource=true">mistralai/codestral-22b-v0.1</a>.</li> <li>Added multiple deployment settings with A100 (40G), A100 (80G) and H100 (80G) for select models, such as <a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama3_1">llama3.1</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/meta/model-garden/llama3">llama3</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma2">gemma2</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma">gemma</a>, and <a href="https://console.cloud.google.com/vertex-ai/publishers/mistral-ai/model-garden/mistral-7b">mistral-7b</a>.</li> </ul> </div> <h2 id="July_30_2024" data-text="July 30, 2024" tabindex="-1">July 30, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="544b332c" class="release-announcement"> <p>See the <a href="https://cloud.google.com/vertex-ai/generative-ai/sla">Gemini Online Inference on Vertex AI Service Level Agreement (SLA)</a>.</p> </div> <h2 id="July_24_2024" data-text="July 24, 2024" tabindex="-1">July 24, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2a69f7f8" class="release-announcement"> <p><strong>Mistral AI</strong></p> <p>Managed models from Mistral AI are available on Vertex AI. To use a Mistral AI model on Vertex AI, send a request directly to the Vertex AI API endpoint. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/mistral">Mistral AI models</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="a43cfaa5" class="release-feature"> <p><strong>M123 release</strong></p> <p>The M123 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a bug that caused conflicting permissions with the Jupyter user and google-sudoers.</li> <li>Updated Nvidia drivers to version 550.90.07 to fix vulnerabilities.</li> </ul> </div> <h2 id="July_23_2024" data-text="July 23, 2024" tabindex="-1">July 23, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6df7aa3b" class="release-announcement"> <p><strong>Llama 3.1</strong></p> <p>The Llama 3.1 405B model is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a> on Vertex AI. Llama 3.1 405B provides capabilities from synthetic data generation to model distillation, steerability, math, tool use, multilingual translation, and more. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/llama">Llama models</a>. </p> </div> <h2 id="July_16_2024" data-text="July 16, 2024" tabindex="-1">July 16, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="d6440c6f" class="release-changed"> <p>All Colab Enterprise runtimes are automatically configured with a 100 GiB boot disk in addition to the disk specified in the runtime template. Starting July 16, 2024, the boot disk of a newly created Colab Enterprise runtime automatically defaults to an SSD Persistent Disk. Previously, the boot disk default was a Standard Persistent Disk. </p> <p>Because of this change, default boot disks of Colab Enterprise runtimes are billed as SSD Persistent Disks instead of Standard Persistent Disks. For more information, see <a href="https://cloud.google.com/colab/pricing">Colab Enterprise pricing</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="216ee564" class="release-feature"> <p><strong>M123 release</strong></p> <p>The M123 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Fixed a bug that caused conflicting permissions with the Jupyter user and google-sudoers.</li> <li>Fixed a bug for custom container instances using a disabled root.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="df70bc0b" class="release-feature"> <p><strong>M123 release</strong></p> <p>The M123 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Fixed a bug that caused conflicting permissions with the Jupyter user and google-sudoers.</li> </ul> </div> <h2 id="July_02_2024" data-text="July 02, 2024" tabindex="-1">July 02, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b0eec494" class="release-announcement"> <p>Google's open weight Gemma 2 model is available in Model Garden. For details, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-gemma">Use Gemma open models</a>.</p> </div> <div id="d004c344" class="release-changed"> <p>MaMMUT is now available in <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>. MaMMUT is a vision-encoder and text-decoder model for multimodal tasks such as visual question answering, image-text retrieval, text-image retrieval, and generation of multimodal embeddings.</p> </div> <h2 id="June_28_2024" data-text="June 28, 2024" tabindex="-1">June 28, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="51adb40d" class="release-feature"> <p>The following models have been added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>:</p> <ul> <li>36 Hugging Face embedding models with verified deployment settings such as <a href="https://console.cloud.google.com/vertex-ai/publishers/BAAI/model-garden/bge-m3;action=deploy;hfSource=true">BAAI/bge-m3</a> and <a href="https://console.cloud.google.com/vertex-ai/publishers/intfloat/model-garden/multilingual-e5-large-instruct;action=deploy;hfSource=true">intfloat/multilingual-e5-large-instruct</a>.</li> <li>35 Hugging Face PyTorch models with verified deployment settings such as <a href="https://console.cloud.google.com/vertex-ai/publishers/stabilityai/model-garden/stable-diffusion-2-1;action=deploy;hfSource=true">stabilityai/stable-diffusion-2-1</a>.</li> </ul> <p>For more information, see the <a href="http://console.cloud.google.com/vertex-ai/model-garden;action=deploy;hfSource=true">Hugging Face model deployment</a> in the console.</p> </div> <div id="45e90797" class="release-feature"> <p>Launched <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-hex-llm">Hex-LLM</a> for high-efficiency large language model serving. This performant TPU serving solution is based on XLA and optimized kernels to achieve high throughput and low latency.</p> <p>Hex-LLM uses several parallelism strategies for multiple TPU chips, quantizations, dynamic LoRA, and more. Hex-LLM supports the following dense and sparse LLMs:</p> <ul> <li>Gemma 2B and 7B</li> <li>Gemma 2 9B and 27B</li> <li>Llama 2 7B, 13B and 70B</li> <li>Llama 3 8B and 70B</li> <li>Mistral 7B and Mixtral 8x7B</li> </ul> </div> <div id="525d78f8" class="release-changed"> <ul> <li>Updated Docker images in <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_llama3_finetuning.ipynb">Llama 3 notebooks</a> that are more efficient at tuning.</li> <li>A notebook-based interactive workshop UI was added in <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a> for image generative models such as <a href="https://console.cloud.google.com/vertex-ai/publishers/stability-ai/model-garden/stable-diffusion-xl-base">stable-diffusion-xl-base</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/runwayml/model-garden/stable-diffusion-inpainting">image inpainting</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/lllyasviel/model-garden/control-net">controlnet</a>. You can find these models from the <strong>Open Notebook</strong> list.</li> <li>Colab Notebooks for frequently used models in Model Garden have been revised with no-code or low-code implementations to improve accessibility and user experience.</li> </ul> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c292c71c" class="release-announcement"> <p>Vertex AI custom training on TPU VMs support <a href="https://cloud.google.com/vertex-ai/docs/general/cmek">customer managed encryption keys (CMEK)</a>.</p> </div> <h2 id="June_27_2024" data-text="June 27, 2024" tabindex="-1">June 27, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9f1b212f" class="release-feature"> <p>Context caching is available for Gemini 1.5 Pro. Use context caching to reduce the cost of requests that contain repeat content with high input token counts. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/context-cache/context-cache-overview">Context caching overview</a>.</p> </div> <h2 id="June_25_2024" data-text="June 25, 2024" tabindex="-1">June 25, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="21d76322" class="release-feature"> <p>Controlled generation is available on Gemini 1.5 Pro and supports the JSON schema. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/control-generated-output">Control generated output</a>.</p> </div> <h2 id="June_21_2024" data-text="June 21, 2024" tabindex="-1">June 21, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="972df924" class="release-feature"> <p><strong>M122 release</strong></p> <p>The M122 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Updated Nvidia drivers to version 550.90.07 to fix vulnerabilities.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="78c18b93" class="release-feature"> <p><strong>M122 release</strong></p> <p>The M122 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Updated Nvidia drivers to version 550.90.07 to fix vulnerabilities.</li> </ul> </div> <h2 id="June_20_2024" data-text="June 20, 2024" tabindex="-1">June 20, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4fdb6e26" class="release-announcement"> <p>The Anthropic Claude Sonnet 3.5 is Generally Available. To learn more, view the <a href="https://console.cloud.google.com/vertex-ai/publishers/anthropic/model-garden/claude-3-5-sonnet">Claude Sonnet 3.5 model card</a> in Model Garden.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4077bb1a" class="release-announcement"> <p>Vertex AI custom training supports TPU v5e in <code translate="no" dir="ltr">us-central1</code>. For details, see <a href="https://cloud.google.com/vertex-ai/docs/general/locations">Vertex AI locations</a>.</p> </div> <h2 id="June_18_2024" data-text="June 18, 2024" tabindex="-1">June 18, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f5597041" class="release-deprecated"> <p>Starting on September 15, 2024, you can only customize classification, entity extraction, and sentiment analysis objectives by moving to Vertex AI Gemini prompts and tuning. Training or updating models for Vertex AI AutoML for Text classification, entity extraction, and sentiment analysis objectives will no longer be available. You can continue using existing Vertex AI AutoML Text models until June 15, 2025. For more information about how Gemini offers enhanced user experience through improved prompting capabilities, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/tune-models">Overview of model tuning for Gemini</a>.</p> </div> <h2 id="June_17_2024" data-text="June 17, 2024" tabindex="-1">June 17, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="4632af01" class="release-feature"> <p>You can now use customer-managed encryption keys (CMEK) to protect runtimes in Colab Enterprise. Using CMEK for notebook files isn't currently supported.</p> <p>For more information, see <a href="https://cloud.google.com/colab/docs/cmek">Use customer-managed encryption keys for runtimes</a>.</p> </div> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b8a56248" class="release-changed"> <p>Increased the input token limit for Gemini 1.5 Pro from 1M to 2M. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models">Google models</a>.</p> </div> <h2 id="June_11_2024" data-text="June 11, 2024" tabindex="-1">June 11, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f88ee62d" class="release-changed"> <p><strong>Upload media from Google Drive</strong></p> <p>You can upload media, such as PDF, MP4, WAV, and JPG files from Google Drive, when you send <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/image-understanding">image</a>, <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/video-understanding">video</a>, <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/audio-understanding">audio</a>, and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/document-understanding">document</a> prompt requests.</p> </div> <h2 id="June_10_2024" data-text="June 10, 2024" tabindex="-1">June 10, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="69cd09d6" class="release-feature"> <p>Gemini in Colab Enterprise, which is a product in the Gemini for Google Cloud portfolio, is available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. Gemini in Colab Enterprise helps you write code by suggesting code as you type. You can also use the <strong>Help me code</strong> tool to generate code from a description of what you want.</p> <p>To learn how to enable and activate Gemini in Colab Enterprise features, see <a href="https://cloud.google.com/gemini/docs/colab/set-up-gemini">Set up Gemini in Colab Enterprise</a>.</p> </div> <div id="643a4407" class="release-feature"> <p>The notebook scheduler is now available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. You can schedule a notebook to run immediately one time, or on a recurring schedule.</p> <p>For more information, see <a href="https://cloud.google.com/colab/docs/schedule-notebook-run">Schedule a notebook run</a>.</p> </div> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bbead42a" class="release-feature"> <p><strong>Experiment in the Vertex AI Studio login-free</strong></p> <p>The Vertex AI Studio multi-model prompt designer can be accessed login-free. With this feature, prospective customers can use the Vertex AI Studio to test queries before deciding to sign up and create an account. To learn more about this experience, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/overview#console-experiences">Vertex AI Studio console experiences</a> or to access the console directly go to <a href="https://console.cloud.google.com/vertex-ai/generative/multimodal/create/text?_ga=2.160908414.2102276437.1718032367-729345960.1717709985">Vertex AI Studio</a>.</p> </div> <h2 id="June_07_2024" data-text="June 07, 2024" tabindex="-1">June 07, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="98545a1d" class="release-feature"> <p>You can now create a Vertex AI Workbench instance based on a custom container. This feature is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Only custom containers derived from the Google-provided base container are supported. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/create-custom-container">Create an instance using a custom container</a>.</p> </div> <h2 id="June_03_2024" data-text="June 03, 2024" tabindex="-1">June 03, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="cdd242f1" class="release-feature"> <p>You can now use <a href="https://cloud.google.com/iam/docs/workforce-identity-federation">Workforce Identity Federation</a> with Vertex AI Workbench instances in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. Workforce Identity Federation lets you create and manage Vertex AI Workbench instances with credentials provided by an external identity provider (IdP). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/create-third-party-instance">Create an instance with third party credentials</a>.</p> </div> <h2 id="May_31_2024" data-text="May 31, 2024" tabindex="-1">May 31, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="cccd2519" class="release-changed"> <p><strong>Anthropic Claude 3.0 Opus model</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-claude#claude-opus">Anthropic Claude 3.0 Opus</a> model is <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>. To learn more, see its <a href="https://console.cloud.google.com/vertex-ai/publishers/anthropic/model-garden/claude-3-opus">model card in Model Garden</a>.</p> </div> <div id="85ad9286" class="release-changed"> <p><strong>Generative AI on Vertex AI Regional APIs</strong></p> <p>Generative AI on Vertex AI regional APIs are available in the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/locations">following three regions</a>: <ul> <li><code translate="no" dir="ltr">us-east5</code></li><li><code translate="no" dir="ltr">me-central1</code></li><li><code translate="no" dir="ltr">me-central2</code></li></ul></p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="871b6fc1" class="release-feature"> <p>Model Monitoring v2 is in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>, which centralizes model monitoring configuration and visualization on a model version and enables monitoring models being served outside of Vertex AI. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/model-monitoring/overview">Vertex AI Model Monitoring overview</a>.</p> </div> <div id="3e28b259" class="release-changed"> <p><strong>Vertex AI Regional APIs</strong></p> <p>Vertex AI regional APIs are available in the <a href="https://cloud.google.com/vertex-ai/docs/general/locations#available-regions">following seven regions</a>: <ul> <li><code translate="no" dir="ltr">us-east5</code></li><li><code translate="no" dir="ltr">us-south1</code></li><li><code translate="no" dir="ltr">africa-south1</code></li><li><code translate="no" dir="ltr">europe-southwest1</code></li><li><code translate="no" dir="ltr">europe-west12</code></li><li><code translate="no" dir="ltr">me-central1</code></li><li><code translate="no" dir="ltr">me-central2</code></li></ul></p> </div> <h2 id="May_28_2024" data-text="May 28, 2024" tabindex="-1">May 28, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6f795a5a" class="release-feature"> <p>Gemini models support the <code translate="no" dir="ltr">frequencyPenalty</code> and <code translate="no" dir="ltr">presencePenalty</code> parameters. Use <code translate="no" dir="ltr">frequencyPenalty</code> to control the probability of repeated text in a response. Use <code translate="no" dir="ltr">presencePenalty</code> to control the probability of generating more diverse content. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/gemini#parameters">Gemini model parameters</a>.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f95a1982" class="release-announcement"> <p><strong>Vector Search sparse embeddings and hybrid search in Public preview</strong></p> <p>Vector Search supports sparse embeddings and hybrid search in Public preview. Hybrid search uses both dense and sparse embeddings, which lets you search based on a combination of keyword search and semantic search. For how to format dense, sparse, and hybrid embeddings, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/setup/format-structure">Input data and structure</a>.</p> </div> <h2 id="May_24_2024" data-text="May 24, 2024" tabindex="-1">May 24, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c19f297e" class="release-announcement"> <p>The Gemini 1.5 Pro (<code translate="no" dir="ltr">gemini-1.5-pro-001</code>) and Gemini 1.5 Flash (<code translate="no" dir="ltr">gemini-1.5-flash-001</code>) models are <a href="https://cloud.google.com/products?e=13802955&hl=en#product-launch-stages">Generally Available</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/models">Google models</a>, <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/overview">Overview of the Gemini API</a>, and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/send-multimodal-prompts">Send multimodal prompt requests</a>. </p> </div> <h2 id="May_20_2024" data-text="May 20, 2024" tabindex="-1">May 20, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ef027817" class="release-feature"> <p>The following models have been added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>:</p> <ul> <li><strong>E5</strong>: A text embedding model series that can be served with a GPU or CPU.</li> <li><strong>Instant ID</strong>: An identity preserving text-to-image generation model.</li> <li><strong>Stable Diffusion XL lightning</strong>: A text-to-image generation model that is based on SDXL but requires fewer inference iterations. </li> </ul> <p>To see a list of all available models, see <a href="https://cloud.google.com/vertex-ai/docs/start/explore-models">Explore models in Model Garden</a>.</p> </div> <h2 id="May_17_2024" data-text="May 17, 2024" tabindex="-1">May 17, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="c87ee64d" class="release-feature"> <p><strong>M121 release</strong></p> <p>The M121 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Updated Nvidia drivers to 550.54.15 to fix an issue where Nvidia drivers failed to install on startup after Debian 11 images upgraded kernel to <code translate="no" dir="ltr">linux-image-5.10.0-29-cloud-amd64</code>.</li> <li>The <code translate="no" dir="ltr">linux-headers-cloud-amd64</code> metapackage is now installed for faster driver recompiling on kernel upgrades.</li> <li>TensorFlow 2.6 CPU and GPU images are deprecated. There will be no further updates to these images in future releases.</li> </ul> </div> <div id="2e61be0f" class="release-feature"> <p>The M121 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Updated the R CPU kernel from R 4.3 to R 4.4.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="5b54845b" class="release-feature"> <p><strong>M121 release</strong></p> <p>The M121 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Updated Nvidia drivers to 550.54.15 to fix an issue where Nvidia drivers failed to install on startup after Debian 11 images upgraded kernel to <code translate="no" dir="ltr">linux-image-5.10.0-29-cloud-amd64</code>.</li> <li>The <code translate="no" dir="ltr">linux-headers-cloud-amd64</code> metapackage is now installed for faster driver recompiling on kernel upgrades.</li> </ul> </div> <h2 id="May_14_2024" data-text="May 14, 2024" tabindex="-1">May 14, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d30a0eec" class="release-announcement"> <p><strong>Gemini 1.5 Flash (Preview)</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/send-multimodal-prompts">Gemini 1.5 Flash</a> (<a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versioning#preview-models"><code translate="no" dir="ltr">gemini-1.5-flash-preview-0514</code></a>) is available in <a href="https://cloud.google.com/products?e=13802955&hl=en#product-launch-stages">Preview</a>. Gemini 1.5 Flash is a multimodal model designed for fast, high volume, cost-effective text generation and chat applications. It can analyze text, code, audio, PDF, video, and video with audio.</p> </div> <div id="850bbdee" class="release-feature"> <p><strong>Grounding Gemini with Google Search is GA</strong></p> <p>The Gemini API Grounding with Google Search feature is available in <a href="https://cloud.google.com/products#product-launch-stages">GA</a>. This is available for Gemini 1.0 Pro models. To learn more about model grounding, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/ground-gemini#web-ground-gemini">Grounding with Google Search</a>.</p> </div> <div id="8a4b3a3f" class="release-feature"> <p><strong>Batch prediction support for Gemini</strong></p> <p>Batch prediction is available for Gemini in <a href="https://cloud.google.com/products#product-launch-stages">preview</a>. Available Gemini models include Gemini 1.0 Pro, Gemini 1.5 Pro, and Gemini 1.5 Flash. To get started with batch prediction, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/batch-prediction-gemini">Get batch predictions for Gemini</a>.</p> </div> <div id="bcca8aa1" class="release-announcement"> <p><strong>PaliGemma model</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-gemma">PaliGemma</a> model is available. PaliGemma is a lightweight open model that's part of the Google Gemma model family. It's the Gemma model family's best model option for image captioning tasks and visual question and answering tasks. Gemma models are based on Gemini models and intended to be extended by customers.</p> </div> <div id="90f316a8" class="release-feature"> <p><strong>New stable text embedding models</strong></p> <p>The following text embedding models are available <a href="https://cloud.google.com/products#product-launch-stages">GA</a>:</p> <ul> <li><code translate="no" dir="ltr">text-embedding-004</code></li> <li><code translate="no" dir="ltr">text-multilingual-embedding-002</code></li> </ul> <p>For details on how to use these models, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings/get-text-embeddings">Get text embeddings</a>.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="22874788" class="release-announcement"> <p>Ray on Vertex AI is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>) and includes the following updates:</p> <ul> <li>Ray version 2.9.3 and Python 3.10 are supported. For information about Ray image support policies, see <a href="https://cloud.google.com/vertex-ai/docs/supported-frameworks-list#ray">Supported versions</a>.</li> <li>VPC peering connection is no longer required if you use public endpoints.</li> <li>Custom images are supported with Ray on Vertex AI.</li> <li>You can use custom service accounts with Ray on Vertex AI.</li> <li>A Colab template is not automatically created when you create a Ray Cluster. Instead, you can connect directly to Ray on Vertex AI clusters from <a href="https://cloud.google.com/vertex-ai/docs/open-source/ray-on-vertex-ai/develop-application">Colab Enterprise's side panel</a>.</li> </ul> </div> <div id="123c6d16" class="release-changed"> <p>For Ray on Vertex AI, Ray version 2.4 is no longer supported. Migrate your code to support Ray 2.9.3 or later and then delete Ray clusters that are running 2.4.</p> </div> <h2 id="April_30_2024" data-text="April 30, 2024" tabindex="-1">April 30, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="78dd04ce" class="release-feature"> <p>Vertex AI custom training supports TPU v5e. For details, see <a href="https://cloud.google.com/vertex-ai/docs/training/training-with-tpu-vm">Training with TPU accelerators</a>.</p> </div> <h2 id="April_29_2024" data-text="April 29, 2024" tabindex="-1">April 29, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="04beed9d" class="release-feature"> <p><strong>M120 release</strong></p> <p>The M120 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Minor bug fixes for the <code translate="no" dir="ltr">libcurl</code> package.</li> </ul> </div> <h2 id="April_25_2024" data-text="April 25, 2024" tabindex="-1">April 25, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="02cf1371" class="release-feature"> <p><strong>M120 release</strong></p> <p>The M120 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Upgraded TensorFlow 2.15 user-managed notebooks to TensorFlow 2.15.1.</li> <li>Minor bug fixes for the <code translate="no" dir="ltr">libcurl</code> package.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="46316592" class="release-feature"> <p><strong>M120 release</strong></p> <p>The M120 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Minor bug fixes for the <code translate="no" dir="ltr">libcurl</code> package.</li> </ul> </div> <h2 id="April_18_2024" data-text="April 18, 2024" tabindex="-1">April 18, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="365e290a" class="release-feature"> <p>Meta's open weight <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-llama">Llama 3 model</a> is available in the Vertex AI Model Garden. </p> </div> <h2 id="April_15_2024" data-text="April 15, 2024" tabindex="-1">April 15, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="cf1b1876" class="release-announcement"> <p><a href="https://cloud.google.com/vertex-ai/docs/training/persistent-resource-overview">Persistent resource</a> for Vertex AI custom training is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <div id="7b0a1cc2" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p>The following features of Vertex AI Feature Store are now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>):</p> <ul> <li><p><strong>Optimized online serving</strong>: Serve features at ultra-low latencies. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/online-serving-types#optimized_serving">Optimized online serving</a>.</p></li> <li><p><strong>Search using embeddings</strong>: Perform vector similarity searches to retrieve semantically similar or related features for real-time serving. You can search using embeddings if your online store is configured to support embeddings. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/embeddings-search">Search using embeddings</a>.</p></li> <li><p><strong>Feature view sync</strong>: Refresh or synchronize the feature data in a feature view within an online store from the feature data source in BigQuery. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/sync-data">Sync feature data to online store</a>.</p></li> </ul> </div> <h2 id="April_11_2024" data-text="April 11, 2024" tabindex="-1">April 11, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="55e0b1ea" class="release-announcement"> <p><strong>Anthropic Claude 3.0 Opus model</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/use-claude#claude-opus">Anthropic Claude 3.0 Opus</a> model is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. The Claude 3.0 Opus model is an Anthropic partner model that you can use with Vertex AI. It's the most capable of the Anthropic models at performing complex tasks quickly. To learn more, see its <a href="https://console.cloud.google.com/vertex-ai/publishers/anthropic/model-garden/claude-3-opus">model card in Model Garden</a>.</p> </div> <h2 id="April_09_2024" data-text="April 09, 2024" tabindex="-1">April 09, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ce14a539" class="release-feature"> <p><strong>New Imagen on Vertex AI image generation model and features</strong></p> <p>The 006 version of the Imagen 2 image generation model (<code translate="no" dir="ltr">imagegeneration@006</code>) is now available. This model offers the following additional features:</p> <ul> <li>Additional aspect ratios (1:1, 3:4, 4:3, 9:16, 16:9)</li> <li>Digital watermark (<a href="https://deepmind.google/technologies/synthid/">SynthID</a>) enabled by default</li> <li>Watermark verification*</li> <li>New user-configurable safety features (safety setting, person/face setting)</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/model-versioning">Model versions</a> and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/generate-images">Generate images using text prompts</a>.</p> <p>* The <code translate="no" dir="ltr">seed</code> field can't be used while digital watermark is enabled.</p> </div> <div id="bd0cc758" class="release-feature"> <p><strong>New Imagen on Vertex AI image editing model and features</strong></p> <p>The 006 version of the Imagen 2 image editing model (<code translate="no" dir="ltr">imagegeneration@006</code>) is now available. This model offers the following additional features:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-inpainting">Inpainting</a> - Add or remove content from a masked area of an image</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-outpainting">Outpainting</a> - Expand a masked area of an image</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/edit-product-image">Product image editing</a> - Identify and maintain a primary product while changing the background or product position</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/image/model-versioning">Model versions</a>.</p> </div> <div id="26758d96" class="release-changed"> <p><strong>Change in Imagen image generation version 006 (<code translate="no" dir="ltr">imagegeneration@006</code>) <code translate="no" dir="ltr">seed</code> field behavior</strong></p> <p>For the new Imagen image generation model version 006 (<code translate="no" dir="ltr">imagegeneration@006</code>) the <code translate="no" dir="ltr">seed</code> field behavior has changed. For the v.006 model a digital watermark is enabled by default for image generation. To be able to use a <code translate="no" dir="ltr">seed</code> value to get deterministic output you must disable digital watermark generation by setting the following <code translate="no" dir="ltr">parameter</code>: <code translate="no" dir="ltr">"addWatermark": false</code>.</p> <p>For more information, see the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/image-generation">Imagen for image generation and editing API reference</a>.</p> </div> <div id="62f754ef" class="release-announcement"> <p><strong>CodeGemma model</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/open-models/use-gemma">CodeGemma</a> model is available. CodeGemma is a lightweight open model that's part of the Google Gemma model family. CodeGemma is the Gemma model family's code generation and code completion offering. Gemma models are based on Gemini models and intended to be extended by customers.</p> </div> <div id="4a0a861d" class="release-feature"> <p><strong>Grounding Gemini and Grounding with Google Search</strong></p> <p>The Gemini API now supports Grounding with Google Search in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Currently available for Gemini 1.0 Pro models.</p> </div> <div id="b3bf4c67" class="release-changed"> <p><strong>Regional APIs</strong></p> <ul> <li>Regional APIs are <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/locations#available-regions">available in 11 new countries</a> for Gemini, Imagen, and embeddings.</li> <li><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/locations">US and EU</a> have machine-learning processing boundaries for the <code translate="no" dir="ltr">gemini-1.0-pro-001</code>, <code translate="no" dir="ltr">gemini-1.0-pro-002</code>, <code translate="no" dir="ltr">gemini-1.0-pro-vision-001</code>, and <code translate="no" dir="ltr">imagegeneration@005</code> models.</li> </ul> </div> <div id="88bdd849" class="release-feature"> <p><strong>Generative AI on Vertex AI security control update</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/security-controls">Security controls</a> are available for the online prediction feature for Gemini 1.0 Pro and Gemini 1.0 Pro Vision.</p> </div> <div id="a7053d7a" class="release-announcement"> <p><strong>Gemini 1.5 Pro (Preview)</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/send-multimodal-prompts">Gemini 1.5 Pro</a> (<a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versioning#preview-models"><code translate="no" dir="ltr">gemini-1.5-pro-preview-0409</code></a>) is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Gemini 1.5 Pro is a multimodal model that analyzes text, code, audio, PDF, video, and video with audio. </p> </div> <div id="08bf1d7e" class="release-feature"> <p><strong>New text embedding models</strong></p> <p>The following text embedding models are now in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> <ul> <li><code translate="no" dir="ltr">text-embedding-preview-0409</code></li> <li><code translate="no" dir="ltr">text-multilingual-embedding-preview-0409</code></li> </ul> <p>When evaluated using the <a href="https://huggingface.co/spaces/mteb/leaderboard">MTEB benchmarks</a>, these models produce better embeddings compared to previous versions. The new models also offer <a href="https://arxiv.org/pdf/2205.13147.pdf">dynamic embedding sizes</a>, which you can use to output smaller embedding dimensions, with minor performance loss, to save on computing and storage costs.</p> <p>For details on how to use these models, refer to the <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/embeddings/get-text-embeddings#latest_models">public documentation</a> and try out our <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/generative_ai/text_embedding_new_api.ipynb">Colab</a>.</p> </div> <div id="832da1a8" class="release-feature"> <p><strong>System instructions</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/send-chat-prompts-gemini#system-instructions">System instructions</a> are supported in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a> by the Gemini 1.0 Pro (stable version <code translate="no" dir="ltr">gemini-1.0-pro-002</code> only) and Gemini 1.5 Pro (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>) multimodal models. Use system instructions to guide model behavior based on your specific needs and use cases. For more information, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/multimodal/send-chat-prompts-gemini#system_instructions_examples">System instructions examples</a>.</p> </div> <div id="f5bad2d6" class="release-feature"> <p><strong>Supervised Tuning for Gemini</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/tune-gemini-overview">Supervised tuning</a> is available for the <code translate="no" dir="ltr">gemini-1.0-pro-002 model</code>.</p> </div> <div id="53f2e910" class="release-feature"> <p><strong>Online Evaluation Service</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/evaluation-overview">Generative AI evaluation</a> supports <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/online-pipeline-services">online evaluation</a> in addition to <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/side-by-side-eval">pipeline evaluation</a>. The list of supported evaluation metrics has also expanded. See <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/model-reference/evaluation">API reference</a> and <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/models/online-pipeline-services">SDK reference</a>. </p> </div> <div id="b30815fa" class="release-changed"> <p><strong>Generative AI Knowledge Base</strong></p> <p>The <a href="https://console.cloud.google.com/products/solutions/details/generative-ai-knowledge-base">Jump Start Solution: Generative AI Knowledge Base</a> demonstrates how to build a simple chatbot with business- and domain-specific knowledge. </p> </div> <div id="1263c711" class="release-feature"> <p><strong>Text translation</strong></p> <p><a href="https://cloud.google.com/vertex-ai/generative-ai/docs/translate/translate-text">Translate text in Vertex AI Studio</a> is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. </p> </div> <div id="28b38a9e" class="release-announcement"> <p><strong>Gemini 1.0 Pro stable version 002</strong></p> <p>The 002 version of the Gemini 1.0 Pro multimodal model (<code translate="no" dir="ltr">gemini-1.0-pro-002</code>) is available. For more information about stable versions of Gemini models, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/learn/model-versioning#stable-version">Gemini model versions and lifecycle</a>.</p> </div> <div id="8c59a20b" class="release-changed"> <p><strong>Vertex AI Studio features and updates</strong> </p> <ul> <li>The Vertex AI Studio supports side-by-side comparison to allow users to compare up to 3 prompts in a side-by-side view.</li> <li>The Vertex AI Studio supports rapid evaluation in console and the ability to upload a ground truth response (or a model response to try to emulate). </li> </ul> <p>To learn more, see <a href="https://cloud.google.com/vertex-ai/generative-ai/docs/start/quickstarts/quickstart">Try your prompts in Vertex AI Studio </a></p> </div> <h2 id="April_02_2024" data-text="April 02, 2024" tabindex="-1">April 02, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1031f964" class="release-feature"> <p><a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a> supports all <a href="https://huggingface.co/docs/text-generation-inference/en/index">Text Generation Inference</a> supported models in <a href="https://huggingface.co/">HuggingFace</a>:</p> <ul> <li>Verified deployment settings for about 400 Hugging Face text generation models (including <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/gemma-7b-it;action=deploy;hfSource=true">google/gemma-7b-it</a>, <a href="https://console.cloud.google.com/vertex-ai/publishers/meta-llama/model-garden/Llama-2-7b-chat-hf;action=deploy;hfSource=true">meta-llama/Llama-2-7b-chat-hf</a>, and <a href="https://console.cloud.google.com/vertex-ai/publishers/mistralai/model-garden/Mistral-7B-v0.1;action=deploy;hfSource=true">mistralai/Mistral-7B-v0.1</a>).</li> <li>Other Hugging Face text generation models have unverified deployment settings that are auto generated.</li> </ul> </div> <h2 id="March_29_2024" data-text="March 29, 2024" tabindex="-1">March 29, 2024</h2> <strong class="release-note-product-title">Generative AI on Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e35d32b4" class="release-changed"> <p>The MedLM-large model infrastructure has been upgraded to improve latency and stability. Responses from the model might be slightly different.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="bff11411" class="release-fixed"> <p><strong>M119 release</strong></p> <p>The M119 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Fixed an issue wherein Dataproc extensions caused JupyterLab to crash when remote kernels weren't available.</li> </ul> </div> <h2 id="March_18_2024" data-text="March 18, 2024" tabindex="-1">March 18, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fed2653e" class="release-changed"> <p><strong><a href="https://cloud.google.com/vertex-ai/docs/vector-search/overview">Vector Search</a> heuristics-based compaction</strong></p> <p>Vector Search uses heuristics-based metrics assess whether to trigger compaction. This prevents unnecessary compaction, and thus reduces cost. For general information about compaction, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/update-rebuild-index#index-compaction">Compaction</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="d1622bca" class="release-feature"> <p><strong>M118 release</strong></p> <p>The M118 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>PyTorch 2.1.0 with CUDA 12.1 and Python 3.10 user-managed notebooks instances are now available.</li> <li>PyTorch 2.2.0 with CUDA 12.1 and Python 3.10 user-managed notebooks instances are now available.</li> <li>Updated Nvidia drivers of older user-managed notebooks images to R535.</li> </ul> </div> <div id="c9320c67" class="release-feature"> <p>The M118 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Updated Nvidia drivers to R535, which fixed a bug where the latest PyTorch 2.0 kernel didn't work due to outdated drivers.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="4d7083ce" class="release-feature"> <p><strong>M118 release</strong></p> <p>The M118 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Updated Nvidia drivers to R535.</li> </ul> </div> <h2 id="March_08_2024" data-text="March 08, 2024" tabindex="-1">March 08, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="479fa62f" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p>The following features of Vertex AI Feature Store are now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>:</p> <ul> <li><p><strong>Integration of Vertex AI Feature Store with Dataplex:</strong> Online store instances, feature views, and feature groups are now automatically registered as data assets in Data Catalog, a Dataplex feature that catalogs metadata from these resources. You can use the metadata search capability of Dataplex to search for and view the metadata of these resources. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/search-feature-metadata">Search for resource metadata in Data Catalog</a>.</p></li> <li><p><strong>Service account configuration for feature views:</strong> You can configure a feature view to use a dedicated service account. By default, every feature view uses the service account configured for your project. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/create-featureview#serviceaccount">Configure the service account for a feature view</a>.</p></li> <li><p><strong>Multiple entity IDs for a feature view:</strong> While creating or updating a feature view, you can specify multiple entity ID columns. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/create-featureview#create_featureview_bq_embedding">Create a feature view</a>.</p></li> </ul> </div> <h2 id="March_05_2024" data-text="March 05, 2024" tabindex="-1">March 05, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="5dacda54" class="release-feature"> <p><strong>Create an empty index with Vector Search</strong></p> <p>You can create an empty index in Vector Search for batch and for streaming. No embedding data is required at index creation time, which enables faster startup time. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/create-manage-index">Manage indexes</a>. </p> </div> <h2 id="March_04_2024" data-text="March 04, 2024" tabindex="-1">March 04, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2e9bb964" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">A3 machine types</a> to serve predictions.</p> </div> <h2 id="February_29_2024" data-text="February 29, 2024" tabindex="-1">February 29, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ffea8dfa" class="release-feature"> <p><strong><a href="https://cloud.google.com/vertex-ai/docs/vector-search/overview">Vector Search</a> feature launch</strong></p> <p><strong>Update streaming index metadata</strong>: With this launch, you can directly update restricts and numeric restricts of data points inside <code translate="no" dir="ltr">StreamUpdate</code> indexes without the compaction cost of a full update. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/update-rebuild-index#update-metadata">Update dynamic metadata</a>. </p> </div> <h2 id="February_28_2024" data-text="February 28, 2024" tabindex="-1">February 28, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="57249fa9" class="release-feature"> <p><strong>M117 release</strong></p> <p>The M117 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Removed the Cloud Storage browser in the left side pane in favor of the existing <strong>Mount shared storage</strong> button.</li> </ul> </div> <h2 id="February_27_2024" data-text="February 27, 2024" tabindex="-1">February 27, 2024</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="909f8d6b" class="release-feature"> <p>VPC Service Controls has <a href="https://cloud.google.com/products#product-launch-stages">general availability</a> support in Colab Enterprise.</p> <p>For more information, see <a href="https://cloud.google.com/colab/docs/service-controls">Use VPC Service Controls</a>.</p> </div> <h2 id="February_26_2024" data-text="February 26, 2024" tabindex="-1">February 26, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="48e9c8e1" class="release-feature"> <p>Structured logging support for Vertex AI custom training. For details, see <a href="https://cloud.google.com/vertex-ai/docs/training/code-requirements#write_code_to_return_container_logs">Write code to return container logs</a>.</p> </div> <div id="20a4fa24" class="release-feature"> <p><strong>Ground Multimodal Models</strong></p> <p><p>Model grounding for <code translate="no" dir="ltr">gemini-pro</code> is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Use grounding to connect the <code translate="no" dir="ltr">gemini-pro</code> model to unstructured text data stores in Vertex AI Search. Grounding lets models access and use the information in the data repositories to generate more enhanced and nuanced responses. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/ground-gemini.md">Ground multimodal models</a>.</p> </div> <h2 id="February_21_2024" data-text="February 21, 2024" tabindex="-1">February 21, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="5184bbbd" class="release-feature"> <p><strong>Gemma open models are available</strong></p> <p>Gemma models, a family of lightweight, open models built from the same research and technology used to create the Gemini models, are available to run on your hardware, mobile devices, or hosted services. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/open-models/use-gemma">Use Gemma open models</a> and the <a href="https://console.cloud.google.com/vertex-ai/publishers/google/model-garden/335">Gemma Model Garden card</a>.</p> </div> <h2 id="February_15_2024" data-text="February 15, 2024" tabindex="-1">February 15, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="3ff889bc" class="release-changed"> <p>The Vertex AI Gemini 1.0 Pro and Gemini 1.0 Pro Vision multimodal language models are generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). They have also been made available in the following regions: europe-west1, europe-west2, europe-west3, europe-west4, and europe-west9.</p> <p>For more information, see the following topics:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/overview">Overview of multimodal models</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/design-multimodal-prompts">Multimodal prompt design</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/model-reference/gemini">Vertex AI Gemini API reference</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/sdk-for-gemini/gemini-sdk-overview">Gemini Python SDK reference</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/migrate/migrate-palm-to-gemini">Migrate from PaLM API to Gemini API</a></li> </ul> </div> <h2 id="February_09_2024" data-text="February 09, 2024" tabindex="-1">February 09, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bf1a2874" class="release-feature"> <p><strong>Multimodal embeddings video support is generally available</strong></p> <p>Embeddings for video data is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>) using the multimodal embedding model (<code translate="no" dir="ltr">multimodalembedding</code>). For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/embeddings/get-multimodal-embeddings#vid-img-txt-request">product documentation</a>.</p> <p>This features incurs pricing based on the mode you use. For more information, see <a href="https://cloud.google.com/vertex-ai/pricing#generative_ai_models">pricing</a>.</p> </div> <h2 id="February_08_2024" data-text="February 08, 2024" tabindex="-1">February 08, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="8368a140" class="release-feature"> <p><strong>M116 release</strong></p> <p>The M116 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Updated custom container user-managed notebooks to use NVIDIA driver version 535.104.05.</li> <li>Fixed bugs in custom container user-managed notebooks where GPUs either wouldn't attach to the container properly, or detached after some time.</li> </ul> </div> <div id="c278b5b7" class="release-feature"> <p>The M116 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a bug (present in versions M113 through M115) that prevented new local kernels from being usable.</li> </ul> </div> <h2 id="February_07_2024" data-text="February 07, 2024" tabindex="-1">February 07, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a8087678" class="release-feature"> <p>The following models have been added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>:</p> <ul> <li>Stable Diffusion XL LCM: The Latent Consistency Model (LCM) enhances text-to-image generation in Latent Diffusion Models by enabling faster and high-quality image creation with fewer steps.</li> <li>LLaVA 1.5: Deploy LLaVA 1.5 models.</li> <li>PyTorch-ZipNeRF: The Pytorch-ZipNeRF model is a state-of-the-art implementation of the ZipNeRF algorithm in the Pytorch framework, designed for efficient and accurate 3D reconstruction from 2D images.</li> <li>LLaMA 2 (Quantized): A quantized version of Meta's Llama 2 models.</li> <li>WizardLM: WizardLM is a large language model (LLM) developed by Microsoft, fine-tuned on complex instructions by adapting the Evol-Instruct method.</li> <li>WizardCoder: WizardCoder is a large language model (LLM) developed by Microsoft, fine-tuned on complex instructions by adapting the Evol-Instruct method to the domain of code.</li> <li>AutoGluon: With AutoGluon you can train and deploy high-accuracy machine learning and deep learning models for tabular data.</li> <li>Lama (Large mask inpainting): Use Large Mask Inpainting with fast Fourier convolutions (FFCs), a high receptive field perceptual loss, and large training masks for resolution-robust image inpainting.</li> </ul> </div> <div id="b8a79de7" class="release-changed"> <p>The following changes have been made to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>:</p> <ul> <li>Added one-click tuning button, and dedicated deployment, tuning, quantization, and evaluation notebooks for Llama 2.</li> <li>Added one-click deployment button for more than 20 models with pre-trained OSS artifacts, including <code translate="no" dir="ltr">Salesforce/blip-image-captioning-base</code> and <code translate="no" dir="ltr">timbrooks/instruct-pix2pix</code>.</li> <li>Supported CodeLlaMA70b with notebooks and the one-click deployment button.</li> <li>Added tuning notebooks for Mistral models.</li> <li>Added serving notebooks for Stable Video Diffusion Img2Vid XT. These notebooks are used for research purposes.</li> </ul> </div> <h2 id="February_05_2024" data-text="February 05, 2024" tabindex="-1">February 05, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="02c3a783" class="release-feature"> <p><strong>Query an index from the Vector Search console</strong></p> <p>Vector Search has launched an improved console experience for querying both private and public deployed indexes, now available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. From the console, you can create an index and endpoint, deploy the index to the endpoint, and query the index for nearest neighbors. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/create-manage-index">Manage indexes</a>. </p> </div> <h2 id="January_29_2024" data-text="January 29, 2024" tabindex="-1">January 29, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e921781c" class="release-feature"> <p><strong>Vertex Prediction</strong></p> <p>You can now customize more deployment parameters when uploading your models, such as shared memory allocation and custom startup and readiness probes. These parameters may be useful when deploying LLMs.</p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/predictions/deploy-genai">Deploy generative AI models</a>, <a href="https://cloud.google.com/vertex-ai/docs/predictions/custom-container-requirements">Custom container requirements for prediction</a>, and <a href="https://cloud.google.com/vertex-ai/docs/reference/rest/v1/ModelContainerSpec"><code translate="no" dir="ltr">ModelContainerSpec</code></a>.</p> </div> <h2 id="January_19_2024" data-text="January 19, 2024" tabindex="-1">January 19, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="9fa3bb89" class="release-feature"> <p><strong>M115 release</strong></p> <p>The M115 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Added support for TensorFlow 2.15 with Python 3.10 on Debian 11.</li> <li>Added support for TensorFlow 2.14 with Python 3.10 on Debian 11.</li> </ul> </div> <div id="c83ed816" class="release-feature"> <p>The M115 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed the BigQuery connector within PySpark containers.</li> </ul> </div> <div class="release-note-product-version-title">v2</div> <div id="7e67bc71" class="release-feature"> <p><strong>M115 release</strong></p> <p>The M115 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Added support for <code translate="no" dir="ltr">venv</code> kernels.</li> </ul> </div> <h2 id="January_16_2024" data-text="January 16, 2024" tabindex="-1">January 16, 2024</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="5df59fcb" class="release-deprecated"> <p>Vertex AI Workbench managed notebooks is <a href="https://cloud.google.com/vertex-ai/docs/deprecations">deprecated</a>. On January 30, 2025, support for managed notebooks will end and the ability to create managed notebooks instances will be removed. Existing instances will continue to function but patches, updates, and upgrades won't be available. To continue using Vertex AI Workbench, you can <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/migrate-to-instances">migrate your managed notebooks instances to Vertex AI Workbench instances</a>.</p> </div> <div id="665f03de" class="release-deprecated"> <p>Vertex AI Workbench user-managed notebooks is <a href="https://cloud.google.com/vertex-ai/docs/deprecations">deprecated</a>. On January 30, 2025, support for user-managed notebooks will end and the ability to create user-managed notebooks instances will be removed. Existing instances will continue to function but patches, updates, and upgrades won't be available. To continue using Vertex AI Workbench, you can <a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/migrate-to-instances">migrate your user-managed notebooks instances to Vertex AI Workbench instances</a>.</p> </div> <h2 id="January_12_2024" data-text="January 12, 2024" tabindex="-1">January 12, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6ee3f9d7" class="release-feature"> <p>Model tuning for the <code translate="no" dir="ltr">textembedding-gecko</code> and <code translate="no" dir="ltr">textembedding-gecko-multilingual</code> models is available in <a href="https://cloud.google.com/products#product-launch-stages">GA</a>. You can use supervised fine-tuning to tune the <code translate="no" dir="ltr">textembedding-gecko</code> model. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-embeddings">Tune text embeddings</a>.</p> </div> <h2 id="January_08_2024" data-text="January 08, 2024" tabindex="-1">January 08, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d89c87ca" class="release-changed"> <p><strong>AutoSxS evaluates LLMs side by side</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/side-by-side-eval">automatic side-by-side (AutoSxS) evaluation tool</a> is available in Preview to A/B test the performance of your LLMs or pre-generated predictions. It's comparable to human evaluators, yet faster, available on-demand, and more cost-efficient.</p> </div> <h2 id="January_05_2024" data-text="January 05, 2024" tabindex="-1">January 05, 2024</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4784c468" class="release-changed"> <p><strong>Generative AI on Vertex AI regional expansion</strong></p> <p>Generative AI on Vertex AI features for Batch Prediction and Model Evaluation are available in <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/locations#available-regions">12 additional Google Cloud regions</a>.</p> </div> <h2 id="December_29_2023" data-text="December 29, 2023" tabindex="-1">December 29, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="31003022" class="release-changed"> <p><strong>Vertex AI regional expansion</strong></p> <p>Vertex AI features for AutoML Forecasting, AutoML Tabular, Batch Prediction, Online Prediction, Pipelines, Training, Vector Search, and Vizier are available in <a href="https://cloud.google.com/vertex-ai/docs/general/locations#feature-availability">10 additional Google Cloud regions</a>.</p> </div> <h2 id="December_27_2023" data-text="December 27, 2023" tabindex="-1">December 27, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a1a998e8" class="release-changed"> <p><strong>Vertex Prediction</strong></p> <p>Quota for <em>Custom model serving</em> is now calculated based on your <a href="https://cloud.google.com/vertex-ai/docs/general/deployment#scaling">deployed model's</a> real-time usage of compute resources.</p> <p>Previously, compute resources, such as the number of Nvidia A100 GPUs being used, were deducted from your project's quota based on the deployed model's <a href="https://cloud.google.com/vertex-ai/docs/reference/rest/v1/DedicatedResources"><code translate="no" dir="ltr">maxReplicaCount</code></a>.</p> <p>This change lets you deploy models based on actual compute usage rather than max usage, but it can prevent your deployed models from autoscaling if your quota is exhausted.</p> </div> <h2 id="December_18_2023" data-text="December 18, 2023" tabindex="-1">December 18, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a3fc28af" class="release-announcement"> <p><strong><a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a> updates:</strong></p> <ul> <li>Support for <a href="https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview">hyperparameter tuning</a> and customized datasets for <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_openllama_peft.ipynb">OpenLLaMA models</a> using the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-text-models-supervised#dataset-format">dataset format used by supervised tuning</a> in Vertex AI.</li> <li>Support for GPTQ conversions for <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_falcon_instruct_peft.ipynb">falcon-instruct models</a>.</li> <li>Add <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_xl_lcm.ipynb">Latent Consistent Models</a>, and <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_stable_diffusion_xl_1_0.ipynb">research purpose only SDXL-Turbo models</a> to stable diffusion XL notebooks.</li> <li>Add Mixtral 8x7B models in the <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/community/model_garden/model_garden_pytorch_mistral_deployment.ipynb">Mistral notebook</a>.</li> </ul> </div> <h2 id="December_14_2023" data-text="December 14, 2023" tabindex="-1">December 14, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="7333a86b" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/tpu/docs/v5e-inference">Cloud TPU v5e</a> to serve online predictions. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/predictions/use-tpu">Use Cloud TPUs for online prediction</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="dfc5a76c" class="release-feature"> <p><strong>M114 release</strong></p> <p>The M114 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Starting with this release, Python 3.7 is no longer available.</li> <li>Upgraded R to 4.3 on Debian 11 Python 3.10 instances.</li> <li>Upgraded JupyterLab to 3.6.6.</li> </ul> </div> <div id="d53442e9" class="release-feature"> <p>The M114 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Starting with this release, Python 3.7 is no longer available.</li> <li>Added new Dataproc extension for remote kernels.</li> <li>Upgraded JupyterLab to 3.6.6.</li> <li>Fixed an issue that sometimes prevented users from running or scheduling notebooks using a default kernel.</li> </ul> </div> <h2 id="December_13_2023" data-text="December 13, 2023" tabindex="-1">December 13, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="371ffe6c" class="release-feature"> <p><b>Vertex AI Gemini models</b></p> <p>Vertex AI Gemini Pro and Gemini Pro Vision multimodal language models are available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For more information, see the following topics:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/overview">Overview of multimodal models</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/design-multimodal-prompts">Multimodal prompt design</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/model-reference/gemini">Vertex AI Gemini API reference</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/multimodal/sdk-for-gemini/gemini-sdk-overview">Gemini Python SDK reference</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/migrate/migrate-palm-to-gemini">Migrate from PaLM API to Gemini API</a></li> </ul> </div> <div id="7c2be269" class="release-feature"> <p><strong>Imagen 2 General Availability</strong></p> <p>The 005 version of Imagen's image generation model (<code translate="no" dir="ltr">imagegeneration@005</code>) is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>) for image generation tasks. This model version is now the default for image generation tasks. For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/generate-images">product documentation</a>.</p> <p>For general information about Imagen models and versions, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/model-versioning">Imagen model versions and lifecycle</a>.</p> </div> <h2 id="December_12_2023" data-text="December 12, 2023" tabindex="-1">December 12, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6b333f26" class="release-feature"> <p><strong>Text embedding model 003 (textembedding-gecko@003) available</strong></p> <p>The updated stable version of the text embedding foundation model, <code translate="no" dir="ltr">textembedding-gecko@003</code>, is available. <code translate="no" dir="ltr">textembedding-gecko@003</code> features improved quality compared to the previous stable versions, <code translate="no" dir="ltr">textembedding-gecko@001</code> and <code translate="no" dir="ltr">textembedding-gecko@002</code>. For more information on model versions, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/model-versioning">Model versions and lifecycle</a>.</p> </div> <h2 id="December_08_2023" data-text="December 08, 2023" tabindex="-1">December 08, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ca0a21ae" class="release-changed"> <p><strong>Generative AI on Vertex AI security control update</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/genai-security-controls">Access Transparency (AXT) security control</a> is available for the following features:</p> <ul> <li>Embeddings for Multimodal online prediction</li> <li>Imagen on Vertex AI online prediction</li> <li>Imagen on Vertex AI tuning</li> </ul> </div> <h2 id="December_06_2023" data-text="December 06, 2023" tabindex="-1">December 06, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b70453c1" class="release-feature"> <p>Version <code translate="no" dir="ltr">@002</code> of the models for text, chat, code, and code chat are available. The <code translate="no" dir="ltr">@002</code> model versions include improved prompt responses. The <code translate="no" dir="ltr">@002</code> models are:</p> <ul> <li><code translate="no" dir="ltr">text-bison@002</code></li> <li><code translate="no" dir="ltr">chat-bison@002</code></li> <li><code translate="no" dir="ltr">code-bison@002</code></li> <li><code translate="no" dir="ltr">codechat-bison@002</code></li> </ul> <p>To ensure that you always use the <em>stable</em> model version, specify the model identifier with the version number. For example, <code translate="no" dir="ltr">text-bison@002</code>. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/model-versioning">Model versions and lifecycle</a>.</p> </div> <div id="1bd48c0d" class="release-feature"> <p>Version 2 of the stable version of the Codey code completion foundation model, named <code translate="no" dir="ltr">code-gecko@002</code>, is available. <code translate="no" dir="ltr">code-gecko@002</code> features improved quality and reduced latency compared to the previous stable version, <code translate="no" dir="ltr">code-gecko@001</code>. These improvements can lead to a higher rate of acceptance.</p> </div> <h2 id="December_05_2023" data-text="December 05, 2023" tabindex="-1">December 05, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="3acb3bd5" class="release-feature"> <p><strong>Grounding with Vertex AI Search</strong></p> <p><p>Model grounding is available in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). Use grounding to connect the <code translate="no" dir="ltr">text-bison</code> and <code translate="no" dir="ltr">chat-bison</code> models to unstructured data stores in Vertex AI Search. Grounding lets models access and use the information in the data repositories to generate more enhanced and nuanced responses. For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/grounding/overview">Grounding Overview</a>.</p> </div> <h2 id="December_01_2023" data-text="December 01, 2023" tabindex="-1">December 01, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="7a633430" class="release-changed"> <p>The following Vertex AI <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a> updates are available:</p> <ul> <li>Updated default model deployment settings with L4 GPUs, such as LLaMA2, falcon-instruct, openllama, Stable Diffusion 1.5, 2.1, and XL models.</li> <li>Support for <a href="https://cloud.google.com/vertex-ai/docs/training/hyperparameter-tuning-overview">hyperparameter tuning</a> and customized datasets for LLaMA2 models using the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-text-models-supervised#dataset-format">dataset format used by supervised tuning</a> in Vertex AI.</li> <li>Recommended LoRA and QLoRA settings for large language model tuning in Vertex AI. For details, see <a href="https://cloud.google.com/vertex-ai/docs/model-garden/lora-qlora">LoRA and QLoRA recommendations for LLMs</a>.</li> <li>Support for AWQ and GPTQ conversions for LLaMA2 and OpenLLaMA models.</li> <li>Benchmark reports for <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/community-content/vertex_model_garden/benchmarking_reports/jax_vit_benchmarking_report.md">ViT pytorch and JAX training</a>, <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/community-content/vertex_model_garden/benchmarking_reports/pytorch_openllama_hyperparameter_tuning_benchmark_report.md">OpenLLaMA 3b/7b/13b hyperparameter tuning</a>, and <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/community-content/vertex_model_garden/benchmarking_reports/stable_diffusion_v1-5_benchmarking_report.md">Stable Diffusion 1.5 tuning and serving</a>.</li> </ul> </div> <h2 id="November_30_2023" data-text="November 30, 2023" tabindex="-1">November 30, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fd969397" class="release-feature"> <p>The Unicorn model size for PaLM 2 for Text is generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). The <code translate="no" dir="ltr">text-unicorn</code> model provides improved response quality and reasoning capability compared to the <code translate="no" dir="ltr">text-bison</code> model. For details, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/models">Model information</a>.</p> </div> <div id="e7f1ff24" class="release-announcement"> <p>Vertex AI's integration of model and dataset metadata into Dataplex's Data Catalog service is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Search and discover these assets across projects and regions in Dataplex. Learn more at <a href="https://cloud.google.com/vertex-ai/docs/model-registry/model-data-catalog">Use Data Catalog to search for model and dataset resources</a>.</p> <p><strong>Note</strong>: For datasets of type <code translate="no" dir="ltr">TEXT_PROMPT</code>, navigating in the UI from Data Catalog back to Vertex AI (via the <strong>Open in Vertex AI</strong> button, or using the <strong>Resource URL</strong> link) results in a blank page. This is a known issue and expected to be fixed in the near future. To directly view <code translate="no" dir="ltr">TEXT_PROMPT</code> datasets in Vertex AI, navigate to the <a href="https://console.cloud.google.com/vertex-ai/generative/language/my-prompts"><strong>Generative AI My Prompts</strong></a> tab. </p> </div> <h2 id="November_24_2023" data-text="November 24, 2023" tabindex="-1">November 24, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6491239d" class="release-feature"> <p><strong>ComputeToken API now available in Preview</strong></p> <p>The ComputeToken API is now available in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). You can use this API to get a list of tokens for a given prompt. A token is a way to represent a common sequence of characters found in a text input. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/compute-token">Get a list of tokens</a>.</p> </div> <h2 id="November_17_2023" data-text="November 17, 2023" tabindex="-1">November 17, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bc550f4c" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p>The following features of the new and improved Vertex AI Feature Store are now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>):</p> <ul> <li><p><strong>Feature Registry</strong>: Register your feature data sources in BigQuery by creating feature groups and features. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/create-featuregroup">Create a feature group</a> and <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/create-feature">Create a feature</a>.</p></li> <li><p><strong>Cloud Bigtable online serving</strong>: Serve features from one or more BigQuery data sources. You can set up Cloud Bigtable online serving by <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/create-onlinestore">defining online serving clusters called online store instances</a> and <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/create-featureview">creating feature views</a> within the online store instances.</p></li> </ul> <p>Note that the following features of Vertex AI Feature Store are still in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>:</p> <ul> <li>Serve features at ultra-low latencies with <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/online-serving-types">Optimized online serving</a>.</li> <li><a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/sync-data">Sync data</a> in a feature view within an online store.</li> <li><a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/embeddings-search">Retrieve vector embeddings</a> for real-time serving.</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/overview">About Vertex AI Feature Store</a>.</p> </div> <h2 id="November_16_2023" data-text="November 16, 2023" tabindex="-1">November 16, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="923d6ff2" class="release-feature"> <p><strong>M113 release</strong></p> <p>The M113 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Added the Dataproc JupyterLab plugin to Vertex AI Workbench instances. To get started, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/create-dataproc-enabled">Create a Dataproc-enabled instance</a>.</li> <li>When using an instance's Google Cloud CLI, <code translate="no" dir="ltr">gcloud config</code> is preset with the following defaults: <ul> <li><code translate="no" dir="ltr">project</code> is set to your instance's project.</li> <li>Your compute region is set to your instance's region.</li> <li>Your Dataproc region is set to your instance's region.</li> </ul></li> <li>Fixed an issue that prevented Dataproc kernels from working.</li> <li>Fixed a CORS (cross-origin resource sharing) error.</li> </ul> </div> <div class="release-note-product-version-title">v1</div> <div id="009517d7" class="release-feature"> <p><strong>M113 release</strong></p> <p>The M113 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Miscellaneous bug fixes and improvements in Python 3.10 notebooks.</li> </ul> </div> <h2 id="November_13_2023" data-text="November 13, 2023" tabindex="-1">November 13, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f6647d67" class="release-feature"> <p><strong>Numerical filtering available in Vertex AI Vector Search</strong></p> <p>With <a href="https://cloud.google.com/vertex-ai/docs/vector-search/overview">Vector Search</a> you can restrict results by "filtering" your index results. In addition to filtering by using categorical restrictions, you can now use numeric filtering. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/filtering">Filter vector matches</a>.</p> </div> <h2 id="November_10_2023" data-text="November 10, 2023" tabindex="-1">November 10, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4c307b6f" class="release-feature"> <p><strong>Generative AI on Vertex AI</strong> <br><br> <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/genai-security-controls">Security controls</a> are available for additional Generative AI on Vertex AI features.</p> </div> <h2 id="November_07_2023" data-text="November 07, 2023" tabindex="-1">November 07, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="81e9814e" class="release-announcement"> <p><a href="https://cloud.google.com/vertex-ai/docs/training/training-with-tpu-vm">Training on TPU VMs</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="November_03_2023" data-text="November 03, 2023" tabindex="-1">November 03, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9af89ad5" class="release-feature"> <p>The following models have been added to <a href="https://console.cloud.google.com/vertex-ai/model-garden">Model Garden</a>:</p> <ul> <li>ImageBind: Multimodal embedding model.</li> <li>Vicuna v1.5: LLM finetuned based on llama2.</li> <li>OWL-ViT v2: SoTA Open Vocabulary Object Detection model.</li> <li>DITO: SoTA Open Vocabulary Object Detection model.</li> <li>NLLB: Multi-language translation model.</li> <li>Mistral-7B: SoTA LLM at small size.</li> <li>BioGPT: LLM finetuned for biomedical domain.</li> <li>BiomedCILP: Multimodal foundational model finetuned for biomedical domain.</li> </ul> <p>To see a list of all available models, see <a href="https://cloud.google.com/vertex-ai/docs/start/explore-models">Explore models in Model Garden</a>.</p> </div> <div id="c28e071e" class="release-announcement"> <p><strong>New <code translate="no" dir="ltr">textembedding-gecko</code> and <code translate="no" dir="ltr">textembedding-gecko-multilingual</code> stable model versions</strong></p> <p>The following stable model versions are available in Generative AI on Vertex AI:</p> <ul> <li><code translate="no" dir="ltr">textembedding-gecko@002</code></li> <li><code translate="no" dir="ltr">textembedding-gecko-multilingual@001</code></li> </ul> <p>For more information on model versions, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/model-versioning">Model versions and lifecycle</a>.</p> </div> <div id="cdc42ef2" class="release-changed"> <p><strong>Model Garden</strong></p> <ul> <li>Improved language model serving throughput. For details, see <a href="https://cloud.google.com/blog/products/ai-machine-learning/serve-open-source-llms-on-google-cloud">Serving open source large language models efficiently on Model Garden</a>. Notebooks in the relevant model cards have been updated accordingly.</li> <li>Inference speed up to 2 times faster compared with original implementation for Stable Diffusion 1.5, 2.1, and XL models.</li> <li>Improved the workflow of the <strong>Deploy</strong> button in all supported model cards.</li> <li>Updated notebooks for Llama2, OpenLlama, and Falcon Instruct with suggested machine specs for model serving, and <a href="https://github.com/EleutherAI/lm-evaluation-harness">EleutherAI's evaluation harness</a> dockers for model evaluation.</li> </ul> </div> <h2 id="November_02_2023" data-text="November 02, 2023" tabindex="-1">November 02, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="3bf0fc7d" class="release-changed"> <p><strong>Generative AI support on Vertex AI</strong></p> <p>Generative AI on Vertex AI can be accessed through <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/locations-genai">12 regional APIs</a> in North America, Europe, and Asia. Regional APIs let customers control where data is stored at-rest.</p> </div> <h2 id="October_30_2023" data-text="October 30, 2023" tabindex="-1">October 30, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e6d9e8dc" class="release-security"> <p><a href="https://cloud.google.com/deep-learning-vm/docs/images">Deep Learning VM Images</a> is a set of prepackaged virtual machine images with a deep learning framework that are ready to be run out of the box. Recently, an out-of-bounds write vulnerability was discovered in the <code translate="no" dir="ltr">ReadHuffmanCodes()</code> function in the <code translate="no" dir="ltr">libwebp</code> library. This might impact images that use this library.</p> <p>Google Cloud continuously scans its publicly published images and updates the packages to assure patched distros are included in the latest releases available for customer adoption. Deep Learning VM Images have been updated to ensure that the latest VM images include the patched distros. Customers adopting the latest VM images are not exposed to this vulnerability.</p> <p>For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/security-bulletins#GCP-2023-036">Vertex AI security bulletin</a>.</p> </div> <h2 id="October_17_2023" data-text="October 17, 2023" tabindex="-1">October 17, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="8f9ee812" class="release-feature"> <p><strong>New Vertex AI Vector Search Console</strong> </p> <p>Vector Search has launched a console experience in Google Cloud for creating and deploying indexes, now available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. From the console, you can create indexes, and create public or VPC endpoints for your indexes, and deploy. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/create-manage-index">Manage indexes</a>.</p> </div> <div id="e806a1d6" class="release-announcement"> <p><strong>Vertex AI Vector Search Improvements</strong> </p> <p>Vector Search has improved the initial index creation process for smaller indexes (<100MB), reducing time to build from about 1 hour to about 5 mins. To get started, see <a href="https://cloud.google.com/vertex-ai/docs/vector-search/quickstart">Vector Search quickstart</a> to create an index. </p> </div> <h2 id="October_11_2023" data-text="October 11, 2023" tabindex="-1">October 11, 2023</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="881b4795" class="release-feature"> <p>Colab Enterprise is now generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). Colab Enterprise combines the popular collaborative features of Colaboratory with the security and compliance capabilities of Google Cloud. Colab Enterprise includes:</p> <ul> <li>Sharing and collaborating functionality, with IAM access control.</li> <li>Google-managed compute and runtime provisioning, with configurable <a href="https://cloud.google.com/colab/docs/create-runtime-template">runtime templates</a>.</li> <li>Integrations with Vertex AI and <a href="https://cloud.google.com/bigquery/docs/explore-data-colab">BigQuery</a>.</li> <li>Inline code completion with <a href="https://cloud.google.com/duet-ai/docs/overview">Duet AI</a> (<a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>) assistance.</li> <li>End-user credential authentication for running your notebook code.</li> <li><a href="https://cloud.google.com/colab/docs/idle-shutdown">Idle shutdown</a> for runtimes (<a href="https://cloud.google.com/products/#product-launch-stages">Experimental</a>).</li> </ul> <p>To get started, see <a href="https://cloud.google.com/colab/docs/introduction">Introduction to Colab Enterprise</a> or <a href="https://cloud.google.com/colab/docs/create-console-quickstart">create a notebook</a> and start coding.</p> </div> <h2 id="October_10_2023" data-text="October 10, 2023" tabindex="-1">October 10, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="02d1df80" class="release-feature"> <p><strong>M112 release</strong></p> <p>The M112 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Miscellaneous bug fixes and improvements.</li> </ul> </div> <h2 id="October_05_2023" data-text="October 05, 2023" tabindex="-1">October 05, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="aebec29a" class="release-feature"> <p><strong>Ray on Vertex AI is now available in Preview</strong></p> <p><a href="https://docs.ray.io/en/latest/ray-overview/index.html" class="external">Ray</a> is an open-source framework for scaling AI and Python applications. Ray provides the infrastructure to perform distributed computing and parallel processing for your machine learning workflow.</p> <p>You can now create Ray clusters and develop your Ray applications on Vertex AI. This feature is in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/open-source/ray-on-vertex-ai/overview">Ray on Vertex AI overview</a>.</p> </div> <h2 id="October_04_2023" data-text="October 04, 2023" tabindex="-1">October 04, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e1ba2be8" class="release-feature"> <p><strong>Model tuning for the <code translate="no" dir="ltr">textembedding-gecko</code> model is now available in Preview</strong></p> <p>You can now use supervised fine-tuning to tune the <code translate="no" dir="ltr">textembedding-gecko</code> model. This feature is in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-embeddings">Tune text embeddings</a>.</p> </div> <div id="7262951d" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">C3 machine types</a> to serve predictions.</p> </div> <div id="76ba4a4c" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p>The new and improved Vertex AI Feature Store is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. With the new Vertex AI Feature Store you can streamline your feature management in the following ways:</p> <ul> <li><p>Store and maintain your offline feature data in BigQuery, taking advantage of the data management capabilities of BigQuery. In the new Vertex AI Feature Store, BigQuery serves as the offline store. You don't need to copy or import feature data to an offline store in Vertex AI.</p></li> <li><p>Register your feature data sources in BigQuery by creating feature groups and features.</p></li> <li><p>Define online serving clusters called online store instances; and then serve features from one or more BigQuery data sources, by aggregating them in a feature view within an online store instance. Use Optimized online serving for ultra-low latency needs and Cloud Bigtable online serving for high data volumes.</p></li> <li><p>Retrieve vector embeddings stored in BigQuery for real-time serving.</p></li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/latest/overview">About Vertex AI Feature Store</a>.</p> </div> <h2 id="October_03_2023" data-text="October 03, 2023" tabindex="-1">October 03, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9c23e243" class="release-security"> <p>TorchServe is used to host PyTorch machine learning models for online prediction. Vertex AI provides pre-built PyTorch model serving containers which depend on TorchServe. Vulnerabilities were recently discovered in TorchServe which would allow an attacker to take control of a TorchServe deployment if its model management API is exposed. Customers with PyTorch models deployed to Vertex AI online prediction are not affected by these vulnerabilities, since Vertex AI does not expose TorchServe's model management API. Customers using TorchServe outside of Vertex AI should take precautions to ensure their deployments are set up securely.</p> <p>For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/security-bulletins#gcp-2023-029">Vertex AI security bulletin</a>.</p> </div> <h2 id="September_25_2023" data-text="September 25, 2023" tabindex="-1">September 25, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="a1433e7d" class="release-feature"> <p>Vertex AI Workbench instances are now generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). Vertex AI Workbench instances combine features from managed notebooks and user-managed notebooks to provide a robust data science solution. Supported features include:</p> <ul> <li>Idle timeout</li> <li>BigQuery and Cloud Storage integrations</li> <li>End-user and service account authentication</li> <li>VPC Service Controls</li> <li>Customer managed encryption keys (CMEK) and Cloud External Key Manager (Cloud EKM)</li> <li>Health status monitoring</li> <li>Scheduled notebook runs</li> <li>Dataproc integration</li> </ul> <p>To get started, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/introduction">Introduction to Vertex AI Workbench instances</a>.</p> </div> <h2 id="September_18_2023" data-text="September 18, 2023" tabindex="-1">September 18, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="d2cc8103" class="release-deprecated"> <p>Debian 10 and Python 3.7 images have reached their end of patch and support life for Vertex AI Workbench managed notebooks and user-managed notebooks. Debian 11 and Python 3.10 images are available.</p> </div> <h2 id="September_14_2023" data-text="September 14, 2023" tabindex="-1">September 14, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1 & v2</div> <div id="c9db27e3" class="release-feature"> <p><strong>M111 release</strong></p> <p>The M111 release of Vertex AI Workbench instances includes the following:</p> <ul> <li>Miscellaneous software updates.</li> </ul> </div> <div id="d5d2111a" class="release-feature"> <p>The M111 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>PyTorch 2.0 user-managed notebooks instances now include PyTorch XLA 2.0.</li> <li>Miscellaneous software updates.</li> </ul> </div> <div id="09f87407" class="release-feature"> <p>The M111 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Miscellaneous software updates.</li> </ul> </div> <h2 id="September_08_2023" data-text="September 08, 2023" tabindex="-1">September 08, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a2c8f784" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">A2 Ultra machines</a> to serve predictions in <code translate="no" dir="ltr">us-central1</code>, <code translate="no" dir="ltr">us-east4</code>, <code translate="no" dir="ltr">europe-west4</code>, and <code translate="no" dir="ltr">asia-southeast1</code>. Each A2 Ultra machine has a fixed number of NVIDIA A100 80GB GPUs attached.</p> </div> <h2 id="September_06_2023" data-text="September 06, 2023" tabindex="-1">September 06, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9c6b23ba" class="release-changed"> <p><strong>Vertex AI Prediction</strong></p> <p>The following prebuilt containers for prediction have been updated: </p> <ul> <li>tf2-cpu.2-12 </li> <li>tf2-gpu.2-12 </li> <li>tf2-cpu.2-11</li> <li>tf2-gpu.2-11</li> <li>tf2-cpu.2-10</li> <li>tf2-gpu.2-10</li> <li>tf2-cpu.2-9</li> <li>tf2-gpu.2-9</li> <li>tf2-cpu.2-8</li> <li>tf2-gpu.2-8</li> <li>sklearn-cpu.1-2</li> <li>xgboost-cpu.1-7</li> <li>pytorch-cpu.2-0</li> <li>pytorch-gpu.2-0</li> <li>pytorch-cpu.1-13</li> <li>pytorch-gpu.1-13</li> </ul> <p>To update your containers, redeploy your models. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/framework-support-policy">Vertex AI framework support policy</a> and <a href="https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers">Prebuilt containers for prediction</a>.</p> </div> <h2 id="September_01_2023" data-text="September 01, 2023" tabindex="-1">September 01, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="01277038" class="release-changed"> <p><strong>Pricing update</strong></p> <p>The pricing for <code translate="no" dir="ltr">text-bison</code> has been reduced to $0.0005 per 1,000 input and output characters. For details, see <a href="https://cloud.google.com/vertex-ai/pricing">Vertex AI Pricing</a>.</p> </div> <h2 id="August_31_2023" data-text="August 31, 2023" tabindex="-1">August 31, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f8870a8c" class="release-feature"> <p>Experiment management: Google Cloud console now supports visualization of your model's performance changes over steps during training, and shows advanced run comparisons. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/experiments/compare-analyze-runs#console-compare-analyze-runs">Compare and analyze runs: Google Cloud console</a>.</p> </div> <h2 id="August_29_2023" data-text="August 29, 2023" tabindex="-1">August 29, 2023</h2> <strong class="release-note-product-title">Colab Enterprise</strong> <div id="d3b06082" class="release-feature"> <p>Colab Enterprise is now available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. Colab Enterprise combines the popular collaborative features of Colaboratory with the security and compliance capabilities of Google Cloud. Colab Enterprise includes:</p> <ul> <li>Sharing and collaborating functionality, with IAM access control.</li> <li>Google-managed compute and runtime provisioning, with configurable runtime templates.</li> <li>Integrations with Vertex AI and BigQuery.</li> <li>Inline code completion with <a href="https://cloud.google.com/duet-ai/docs/overview">Duet AI</a> assistance.</li> <li>End-user credential authentication for running your notebook code.</li> </ul> <p>To get started, see <a href="https://cloud.google.com/colab/docs/introduction">Introduction to Colab Enterprise</a> or <a href="https://cloud.google.com/colab/docs/create-console-quickstart">create a notebook</a> and start coding.</p> </div> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a01503f1" class="release-feature"> <p><strong>Imagen on Vertex AI now offers the following <a href="https://cloud.google.com/products#product-launch-stages">Generally Available (GA)</a> features:</strong></p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/fine-tune-model">Subject model tuning (standard)</a>*</li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/fine-tune-style">Style model tuning</a>*</li> </ul> <p>* Restricted access feature.</p> <p>For more information about Imagen or how to get access to restricted GA features, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/overview">Imagen on Vertex AI overview</a>.</p> </div> <div id="c244dd8a" class="release-feature"> <p><strong>Stream responses from Generative AI models</strong></p> <p>Generative AI model streaming support is now Generally Available <a href="https://cloud.google.com/products#product-launch-stages">(GA)</a>. After you send a prompt, the model returns response tokens as they're generated instead of waiting for the entire output to be available.</p> <p>Supported models are:</p> <ul> <li><code translate="no" dir="ltr">text-bison</code></li> <li><code translate="no" dir="ltr">chat-bison</code></li> <li><code translate="no" dir="ltr">code-bison</code></li> <li><code translate="no" dir="ltr">codechat-bison</code></li> </ul> <p>To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/streaming">Stream responses from Generative AI models</a>.</p> </div> <div id="d7435ce2" class="release-feature"> <p><strong>New Generative AI support on Vertex AI models and expanded language support</strong></p> <p>Generative AI support on Vertex AI has been updated to include new language model candidates (<em>latest</em> models), language models that support input and output tokens up to 32k, and more supported languages. </p> <p>For details, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/models">Available models</a> and <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/model-versioning">Model versions and lifecycle</a>.</p> </div> <div id="f39b97f4" class="release-feature"> <p><strong>Model tuning for the <code translate="no" dir="ltr">text-bison</code> model is now Generally Available (GA)</strong></p> <p>Tuning the <code translate="no" dir="ltr">text-bison</code> model with supervised fine-tuning (SFT) is now Generally Available <a href="https://cloud.google.com/products#product-launch-stages">(GA)</a> . For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-text-models">Tune text models</a>.</p> </div> <div id="8133f85f" class="release-feature"> <p><strong>Model tuning for the <code translate="no" dir="ltr">chat-bison</code> model is now available in Preview</strong></p> <p>You can now use supervised fine-tuning to tune the <code translate="no" dir="ltr">chat-bison</code> model. This feature is in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-code-models">Tune text models</a>.</p> </div> <div id="2d7b1f1e" class="release-feature"> <p><strong>New embedding model now available in Preview</strong></p> <p>Generative AI support on Vertex AI users can now create embeddings using a new model trained on a wide range of non-English languages in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>).</p> <ul> <li><code translate="no" dir="ltr">textembedding-gecko-multilingual</code></li> </ul> <p>To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/embeddings/get-text-embeddings">Get text embeddings</a>.</p> </div> <div id="2661f139" class="release-feature"> <p><strong>Reinforcement learning from human feedback (RLHF) tuning for <code translate="no" dir="ltr">text-bison</code></strong> The Generative AI text generation foundation model (<code translate="no" dir="ltr">text-bison</code>) now supports RLHF tuning. The RLHF tuning feature is in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-models#genai-rlhf-tuning">Use RLHF model tuning</a>.</p> </div> <div id="711e5f69" class="release-feature"> <p><strong>Vertex AI Codey APIs language support</strong></p> <p>Vertex AI Codey APIs now support additional programming languages. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/code/code-models-overview#supported_coding_languages">Supported coding languages</a></p> </div> <div id="a02fbc62" class="release-feature"> <p><strong>Vertex AI Codey APIs now support supervised fine-tuning (SFT)</strong></p> <p>The code chat (<code translate="no" dir="ltr">codechat-bison</code>) and code generation (<code translate="no" dir="ltr">code-bison</code>) Vertex AI Codey APIs models now support supervised fine-tuning (SFT). The supervised-fine tuning for Vertex AI Codey APIs models feature is in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-code-models">Tune code models</a>.</p> </div> <div id="c5c18f2f" class="release-feature"> <p><strong>Metrics-based model evaluation</strong></p> <p>You can evaluate the performance of foundation models and tuned models against an evaluation dataset for classification, summarization, question answering, and general text generation. This feature is available in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>).</p> <p>To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/evaluate-models">Evaluate model performance</a>.</p> </div> <div id="643990e0" class="release-announcement"> <p><strong>Vertex AI Vector Search</strong> is the new product name for <em>Vertex AI Matching Engine</em>.</p> </div> <div id="c20ffa11" class="release-feature"> <p>Vertex AI Model Registry Models and managed datasets are now synced to Dataplex's Data Catalog service. <a href="https://cloud.google.com/data-catalog/docs/concepts/overview">Data Catalog</a> enables organization-wide search and discovery of data artifacts, while still maintaining IAM boundaries. The sync and search of these assets is available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. For more information, see <a href="https://cloud.google.com/data-catalog/docs/how-to/search">Data Catalog documentation</a>.</p> </div> <div id="5c1bf3c3" class="release-feature"> <p><strong>CountToken API now available in Preview</strong></p> <p>The CountToken API is now available in (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>). You can use this API to get the token count and the number of billable characters for a prompt. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/get-token-count">Get token count</a>.</p> </div> <div id="61a7eef2" class="release-feature"> <p>The Vertex AI Pipelines <strong>Template Gallery</strong> is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). The Template Gallery contains Google-authored pipeline and component templates to bootstrap your MLOps practice. Customize and run the templates as-is or embed them into your own pipelines. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/use-template-gallery">Use a prebuilt template from the Template Gallery</a>.</p> </div> <h2 id="August_28_2023" data-text="August 28, 2023" tabindex="-1">August 28, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b6077112" class="release-feature"> <p>Tabular Workflow for Forecasting is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/forecasting">Tabular Workflow for Forecasting</a>.</p> </div> <h2 id="August_22_2023" data-text="August 22, 2023" tabindex="-1">August 22, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="0be51c6b" class="release-feature"> <p>Vertex AI custom training has launched <a href="https://cloud.google.com/vertex-ai/docs/training/persistent-resource-overview">persistent resources</a> in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. A persistent resource is a long-running cluster of machines that you can use to run custom training jobs. Once created, the persistent resource remains available for future training jobs, so you don't have to wait for compute resources to be provisioned each time you want to train a model. </p> </div> <h2 id="August_18_2023" data-text="August 18, 2023" tabindex="-1">August 18, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="829373bf" class="release-feature"> <p>The Vertex AI Matching Engine public endpoint is now generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). For information about how to get started, see <a href="https://cloud.google.com/vertex-ai/docs/matching-engine/match-eng-setup/match-eng-setup.md">Matching Engine Setup</a>.</p> </div> <h2 id="August_11_2023" data-text="August 11, 2023" tabindex="-1">August 11, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b8465c51" class="release-changed"> <p>Generative AI on Vertex AI supports CMEK, VPC Service Controls, Data Residency, and Access Transparency. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/genai-security-controls">Security controls</a>.</p> </div> <h2 id="August_10_2023" data-text="August 10, 2023" tabindex="-1">August 10, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1 & v2</div> <div id="9a286e3c" class="release-feature"> <p><strong>M110 release</strong></p> <p>The M110 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Added support for TensorFlow 2.13 with Python 3.10 on Debian 11.</li> <li>Added support for TensorFlow 2.8 with Python 3.10 on Debian 11.</li> <li>Miscellaneous software updates.</li> </ul> </div> <div id="7270c6fb" class="release-deprecated"> <p>TensorFlow 2.9 user-managed instances are deprecated.</p> </div> <div id="ee5beb47" class="release-feature"> <p>The M110 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Increased shared memory size to available memory capacity.</li> <li>Added support for Python 3.10 on Debian 11. <ul> <li>Added support for PyTorch 2.0 with Python 3.10.</li> </ul></li> </ul> <aside class="note"><strong>Note:</strong><span> Python 3.7 on Debian 10 images are still available.</span></aside> </div> <h2 id="August_09_2023" data-text="August 09, 2023" tabindex="-1">August 09, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="377d61fd" class="release-feature"> <p><strong>Imagen Multimodal embeddings available in GA</strong></p> <p>Imagen on Vertex AI now offers the following <a href="https://cloud.google.com/products/#product-launch-stages">GA</a> feature:</p> <ul> <li>Multimodal embeddings</li> </ul> <p>This feature incurs different <a href="https://cloud.google.com/vertex-ai/pricing#generative_ai_models">pricing</a> based on if you use image input or text input. For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/embeddings/get-image-embeddings">multimodal embeddings</a> feature page.</p> </div> <h2 id="August_02_2023" data-text="August 02, 2023" tabindex="-1">August 02, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2a0e1b0c" class="release-changed"> <p>Prebuilt containers to perform custom training with <a href="https://cloud.google.com/vertex-ai/docs/training/pre-built-containers#tensorflow">TensorFlow 2.12</a> are now generally available <a href="https://cloud.google.com/products#product-launch-stages">(GA)</a>.</p> </div> <div id="dca1b71f" class="release-changed"> <p>Updated prebuilt images for <a href="https://cloud.google.com/vertex-ai/docs/training/pre-built-containers#tensorflow">Tensorflow 2.11</a> are now available.</p> </div> <h2 id="August_01_2023" data-text="August 01, 2023" tabindex="-1">August 01, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="8533f84c" class="release-announcement"> <p>Vertex AI Tensorboard pricing has changed from a per-user monthly license of $300 per month to $10 GiB per month for storage of your logs. This means no more subscription fees. You only pay for the storage you've used. See the <a href="https://github.com/GoogleCloudPlatform/vertex-ai-samples/blob/main/notebooks/official/experiments/delete_outdated_tensorboard_experiments.ipynb">Vertex AI Tensorboard: Delete Outdated Tensorboard Experiments</a> tutorial for how to manage storage.</p> </div> <div id="0bf616c8" class="release-feature"> <p>The schedules API for Vertex AI Pipelines is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). You can schedule recurring pipeline runs in Vertex AI by specifying a frequency in <a href="https://en.wikipedia.org/wiki/Cron" class="external">cron</a> syntax, and optionally the start time and/or end time. Additionally, you can pause, resume, update, and delete schedules.For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/schedule-pipeline-run">Schedule a pipeline run with scheduler API</a>.</p> </div> <h2 id="July_28_2023" data-text="July 28, 2023" tabindex="-1">July 28, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a872d485" class="release-changed"> <p>The <code translate="no" dir="ltr">learning_rate</code> parameter in generative AI model tuning is now <code translate="no" dir="ltr">learning_rate_multiplier</code>. To use the model's or tuning method's default learning rate, use the default <code translate="no" dir="ltr">learning_rate_multiplier</code> value of <code translate="no" dir="ltr">1.0</code>.</p> <p>If you haven't configured <code translate="no" dir="ltr">learning_rate</code> before, no action is needed. If using <code translate="no" dir="ltr">tuning_method=tune_v2</code> with the v2.0.0 pipeline template (Python SDK v1.28.1+), the recommended learning rate is 0.0002. To convert your custom <code translate="no" dir="ltr">learning_rate</code> to <code translate="no" dir="ltr">learning_rate_multiplier</code>, calculate as follows:</p> <div></div><devsite-code><pre class="devsite-click-to-copy" translate="no" dir="ltr" is-upgraded><code translate="no" dir="ltr">learning_rate_multiplier = custom_learning_rate_value / 0.0002 </code></pre></devsite-code> </div> <h2 id="July_19_2023" data-text="July 19, 2023" tabindex="-1">July 19, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v2</div> <div id="5fc83e99" class="release-feature"> <p>Vertex AI Workbench instances are now available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. Vertex AI Workbench instances combine features from managed notebooks and user-managed notebooks to provide a robust data science solution. Supported features include:</p> <ul> <li>Idle timeout</li> <li>BigQuery and Cloud Storage integrations</li> <li>End-user and service account authentication</li> <li>VPC Service Controls</li> <li>Customer managed encryption keys (CMEK)</li> <li>Health status monitoring</li> <li>Run notebooks on a schedule</li> <li>Dataproc integration</li> </ul> <p>To get started, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/instances/introduction">Introduction to Vertex AI Workbench instances</a>.</p> </div> <h2 id="July_18_2023" data-text="July 18, 2023" tabindex="-1">July 18, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="57a8ea2e" class="release-feature"> <p>Model tuning updates for text-bison:</p> <ul> <li>Upgraded tuning pipeline now offers more efficient tuning and better performance on text-bison.</li> <li>New <code translate="no" dir="ltr">learning_rate</code> parameter lets you adjust the step size at each iteration.</li> </ul> <p>For details, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/models/tune-models">Tune language foundation models</a>.</p> </div> <h2 id="July_17_2023" data-text="July 17, 2023" tabindex="-1">July 17, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a2f1c613" class="release-feature"> <p>Imagen on Vertex AI now offers the following <a href="https://cloud.google.com/products#product-launch-stages">Generally Available (GA)</a> features:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/generate-images">Image generation (text-to-image generation)</a>*</li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/edit-images">Image editing</a>*</li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/image-captioning">Image visual captioning</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/visual-question-answering">Visual Question Answering (VQA)</a></li> </ul> <p>* Restricted access feature.</p> <p>For more information about Imagen or how to get access to restricted GA or Preview features, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/overview">Imagen on Vertex AI overview</a>.</p> </div> <div id="2778b537" class="release-feature"> <p>Imagen now supports human face generation for the following features:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/generate-images">Image generation (text-to-image generation)</a>*</li> <li><a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/edit-images">Image editing</a>*</li> </ul> <p>* Restricted access feature.</p> <p>Human face generation is enabled by default, except for images with children and/or celebrities. For more information, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/image/usage-guidelines">usage guidelines</a>.</p> </div> <div id="46507905" class="release-feature"> <p>The Vertex AI PaLM API has added support for the following languages:</p> <ul> <li>Spanish (es)</li> <li>Korean (ko)</li> <li>Hindi (hi)</li> <li>Chinese (zh)</li> </ul> <p>For the complete list of supported languages, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/models#language_support">Supported languages</a>.</p> </div> <h2 id="July_13_2023" data-text="July 13, 2023" tabindex="-1">July 13, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fee7c7f2" class="release-feature"> <p>Support for <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/text/batch-prediction-genai">batch text (<code translate="no" dir="ltr">text-bison</code>) requests</a> is now available in (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). You can review pricing for the <code translate="no" dir="ltr">chat-bison</code> model at <a href="https://cloud.google.com/vertex-ai/pricing#generative_ai_models">Vertex AI pricing page</a>.</p> </div> <h2 id="July_10_2023" data-text="July 10, 2023" tabindex="-1">July 10, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="38be89ac" class="release-feature"> <p>Support for PaLM 2 for <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/chat/chat-prompts">Chat (<code translate="no" dir="ltr">chat-bison</code>)</a> is now available in (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). You can review pricing for the <code translate="no" dir="ltr">chat-bison</code> model at <a href="https://cloud.google.com/vertex-ai/pricing#generative_ai_models">Vertex AI pricing page</a>.</p> </div> <h2 id="July_07_2023" data-text="July 07, 2023" tabindex="-1">July 07, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="52a67c7f" class="release-announcement"> <p><a href="https://cloud.google.com/vertex-ai/docs/training/cloud-storage-file-system">GCSFuse support for custom training</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="July_06_2023" data-text="July 06, 2023" tabindex="-1">July 06, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b3931342" class="release-feature"> <p><strong>Vertex AI model evaluation</strong> is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>) with the following new Preview features:</p> <ul> <li>Model evaluation with sliced metrics.</li> <li>Model evaluation with fairness and bias metrics.</li> <li>Vision error analysis for AutoML image classification models.</li> </ul> </div> <h2 id="June_30_2023" data-text="June 30, 2023" tabindex="-1">June 30, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="60d40336" class="release-feature"> <p><strong>Vertex Explainable AI</strong></p> <p>Support for <a href="https://cloud.google.com/vertex-ai/docs/explainable-ai/overview#example-based">example-based explanations</a> is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <div id="f1d38a7a" class="release-deprecated"> <p><a href="https://cloud.google.com/vertex-ai/docs/datasets/data-labeling-job">Vertex AI data labeling</a> is deprecated and will no longer be available on Google Cloud after July 1, 2024. For new labeling tasks, you can use <a href="https://cloud.google.com/vertex-ai/docs/datasets/label-using-console">add labels using the Google Cloud console</a> or access data labeling solutions from our partners in the <a href="https://console.cloud.google.com/marketplace/">Google Cloud Console Marketplace</a>, such as Labelbox and Snorkel.</p> </div> <h2 id="June_29_2023" data-text="June 29, 2023" tabindex="-1">June 29, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="13c89df5" class="release-feature"> <p><strong>Vertex AI Codey APIs</strong></p> <p>The <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/code/code-models-overview">Vertex AI Codey APIs</a> are now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Use the Codey APIs to create solutions with code generation, code completion, and code chat. Because the Vertex AI Codey APIs are GA, you incur usage costs if you use them. To learn about pricing, see the <a href="https://cloud.google.com/vertex-ai/pricing#generative_ai_models">Generative AI support on Vertex AI pricing page</a>.</p> <p>The models in this release include:</p> <ul> <li><code translate="no" dir="ltr">code-bison</code> (code generation)</li> <li><code translate="no" dir="ltr">codechat-bison</code> (multi-turn code chat)</li> <li><code translate="no" dir="ltr">code-gecko</code> (code completion)</li> </ul> <p>The maximum tokens for input was increased from 4,096 to 6,144 tokens for <code translate="no" dir="ltr">code-bison</code> and <code translate="no" dir="ltr">codechat-bison</code> to allow longer prompts and chat history. The maximum tokens for output was increased from 1,024 to 2,048 for <code translate="no" dir="ltr">code-bison</code> and <code translate="no" dir="ltr">codechat-bison</code> to allow for longer responses.</p> <p>Additional programming languages are supported. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/code/code-models-overview#supported_coding_languages">Supported coding languages</a>.</p> <p>Several fine-tuning datasets were removed from the <code translate="no" dir="ltr">code-bison</code> and <code translate="no" dir="ltr">codechat-bison</code> models to implement the following improvements:</p> <ul> <li>Excessive chattiness.</li> <li>Artifacting, such as NBSP (non-breaking space) characters.</li> <li>Low quality code responses.</li> </ul> <p>To learn about cloud horizontals, please see <a href="https://cloud.google.com/vertex-ai/docs/general/vertexai-compliance">Vertex AI certifications</a>.</p> </div> <div id="f466c256" class="release-feature"> <p>Vertex AI Pipeline task-level logs are now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>) in Cloud Logging. Additionally, from Cloud Logging you can route pipeline logs to a Pub/Sub sink to power your event-driven architecture. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/logging">View pipeline job logs</a>.</p> </div> <h2 id="June_26_2023" data-text="June 26, 2023" tabindex="-1">June 26, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="da5da712" class="release-feature"> <p><strong>M109 release</strong></p> <p>The M109 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>PyTorch 2.0 with Python 3.10 and CUDA 11.8 user-managed notebooks instances are now available.</li> <li>Miscellaneous software updates.</li> </ul> </div> <div id="14e0a916" class="release-feature"> <p>The M109 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a bug that caused high cpu utilization due to excessive internal diagnostic tool processes.</li> <li>Fixed a bug that was showing incorrect kernel image icons in the Jupyterlab launcher.</li> </ul> </div> <h2 id="June_20_2023" data-text="June 20, 2023" tabindex="-1">June 20, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="90ec8116" class="release-feature"> <p>A100 80GB accelerators are now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>) for custom training jobs in the following regions:</p> <ul> <li>asia-southeast1</li> <li>europe-west4</li> <li>us-central1</li> <li>us-east4</li> </ul> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/general/locations#accelerators">Locations</a>.</p> </div> <div id="07ae38bd" class="release-announcement"> <p>The Google Cloud Pipeline Components (GCPC) SDK v2 is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). GCPC v2 introduces support for the KFP v2 SDK and is fully supported by Vertex AI Pipelines.</p> <p>To learn more about the updates in the latest version of the GCPC SDK, see the <a href="https://google-cloud-pipeline-components.readthedocs.io/en/google-cloud-pipeline-components-2.0.0/release.html">Google Cloud Pipelines Components Release Notes</a>.</p> </div> <div id="e5c6f95f" class="release-announcement"> <p>The Kubeflow Pipelines (KFP) SDK v2 is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). KFP SDK v2 introduces several improvements for authoring pipelines and is fully supported by Vertex AI Pipelines.</p> <p>To learn more about the changes in KFP v2, see the <a href="https://github.com/kubeflow/pipelines/blob/master/sdk/RELEASE.md#200">KFP v2 Release Notes</a> and <a href="https://www.kubeflow.org/docs/components/pipelines/v2/migration/">KFP v2 migration guide</a>.</p> </div> <h2 id="June_15_2023" data-text="June 15, 2023" tabindex="-1">June 15, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ba0eb42e" class="release-feature"> <p>The <code translate="no" dir="ltr">chat-bison@001</code> model has been updated to better follow instructions in the <code translate="no" dir="ltr">context</code> field. For details, on how to create chat prompts for <code translate="no" dir="ltr">chat-bison@001</code>, see <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/chat/chat-prompts">Design chat prompts</a>.</p> </div> <h2 id="June_09_2023" data-text="June 09, 2023" tabindex="-1">June 09, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1e2e1ba6" class="release-announcement"> <p><strong>HIPAA compliance for Generative AI on Vertex AI</strong></p> <p>Generative AI support on Vertex AI now supports HIPAA compliance. The coverage includes components of the <a href="https://cloud.google.com/vertex-ai/docs/start/explore-models">Model Garden</a> and <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/generative-ai-studio">Generative AI Studio</a>.</p> <p>To learn more about Vertex certifications, see <a href="https://cloud.google.com/vertex-ai/docs/general/features">Vertex AI features</a> and <a href="https://cloud.google.com/vertex-ai/docs/general/vertexai-compliance">Vertex AI certifications</a>.</p> </div> <h2 id="June_07_2023" data-text="June 07, 2023" tabindex="-1">June 07, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b88274a9" class="release-feature"> <p><strong>PaLM Text and Embeddings APIs, and Generative AI Studio</strong></p> <p>The Generative AI support on Vertex AI is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). With this feature launch, you can leverage the PaLM API to generate AI models that you can test, tune, and deploy in your AI-powered applications. With the GA of these features, you will incur usage costs if you use the <code translate="no" dir="ltr">text-bison</code> and <code translate="no" dir="ltr">textembedding-gecko</code> PaLM APIs. To learn about pricing, see the <a href="https://cloud.google.com/vertex-ai/pricing#generative_ai_models">Vertex AI pricing page</a>.</p> <p>Features and models in this release include:</p> <ul> <li>PaLM 2 for Text: <code translate="no" dir="ltr">text-bison</code></li> <li>Embedding for Text: <code translate="no" dir="ltr">textembedding-gecko</code></li> <li>Generative AI Studio for Language</li> </ul> <aside class="special"><strong>Important:</strong><span> With this GA launch, standard security and compliance for Vertex AI is not yet available to Generative AI. To learn more, see <a href="https://cloud.google.com/vertex-ai/docs/general/features">Vertex AI features</a> and <a href="https://cloud.google.com/vertex-ai/docs/general/vertexai-compliance">Vertex AI certifications</a>.</span></aside> </div> <div id="371a3290" class="release-feature"> <p><strong>Vertex AI Model Garden</strong></p> <p>The Vertex AI Model Garden is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). The Model Garden is a platform that helps you discover, test, customize, and deploy Vertex AI and select OSS models. These models range from tunable to task-specific - all available on the Model Garden page in the Google Cloud console.</p> <p>To get started, see <a href="https://cloud.google.com/vertex-ai/docs/start/explore-models#create-a-vision-app">Explore AI models and APIs in Model Garden</a>.</p> </div> <div id="429c4adc" class="release-feature"> <p><strong>Vertex AI Codey APIs</strong></p> <p>The Vertex AI Codey APIs are now in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. With the Codey API, code generation, code completion, and code chat APIs can be used from any Google Cloud project without allowlisting. The APIs can be accessed from the <code translate="no" dir="ltr">us-central1</code> region. The Codey APIs can be used in the Generative AI studio or programmatically in REST commands.</p> <p>To get started, see the <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/code/code-models-overview">Code models overview</a>.</p> </div> <h2 id="June_01_2023" data-text="June 01, 2023" tabindex="-1">June 01, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2a5e8cd0" class="release-feature"> <p><strong>Vertex Prediction</strong></p> <p>You can now specify a multi-region BigQuery table as the input or output to a <a href="https://cloud.google.com/vertex-ai/docs/predictions/get-predictions#get_batch_predictions">batch prediction request</a>.</p> </div> <h2 id="May_18_2023" data-text="May 18, 2023" tabindex="-1">May 18, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f72f5373" class="release-feature"> <p><strong>Vertex Prediction</strong></p> <p>You can now co-host models on the same VM from the Google Cloud Console. Previously, this capability was available only from the REST API. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/predictions/model-co-hosting">Share resources across deployments</a>.</p> </div> <h2 id="May_16_2023" data-text="May 16, 2023" tabindex="-1">May 16, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="53fb8022" class="release-feature"> <p>Vertex AI custom training now supports deep integration with Vertex AI Experiments. You can submit training jobs with autologging enabled to automatically log parameters and model performance metrics. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/experiments/run-training-job-experiments">Run training job with experiment tracking</a></p> </div> <div id="4378f258" class="release-feature"> <p>The scheduler API for Vertex AI Pipelines is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. You can schedule recurring pipeline runs in Vertex AI by specifying a frequency, start time (optional), and end time (optional). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/schedule-pipeline-run">Schedule a pipeline run with scheduler API</a>.</p> </div> <h2 id="May_10_2023" data-text="May 10, 2023" tabindex="-1">May 10, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b59d28cd" class="release-feature"> <p><strong>Generative AI Support for Vertex AI</strong></p> <p>Generative AI Support for Vertex AI is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. With this feature launch, you can leverage the Vertex AI PaLM API to generate AI models that you can test, tune, and deploy in your AI-powered applications. </p> <p>Features and models in this release include:</p> <ul> <li>PaLM 2 for Text: text-bison@001</li> <li>PaLM 2 for Chat: chat-bison@001</li> <li>Embedding for Text: textembedding-gecko@001 </li> <li>Generative AI Studio for Language</li> <li>Tuning for PaLM 2</li> <li>Vertex AI SDK v1.25, which includes new features such as TextGenerationModel(text-bison@001), ChatModel(chat-bison@001), TextEmbeddingModel(textembedding-gecko@001)</li> </ul> <p>You can interact with the generative AI features on Vertex AI by using Generative AI Studio in the Google Cloud console, the Vertex AI API, and the Vertex AI SDK for Python. </p> <ul> <li>Learn more about <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/overview">Generative AI Support for Vertex AI</a></li> <li>See an <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/learn/generative-ai-studio">Introduction to Generative AI Studio</a></li> <li>Get started with a <a href="https://cloud.google.com/vertex-ai/docs/generative-ai/start/quickstarts/quickstart">Generative AI Studio quickstart</a></li> </ul> </div> <div id="7005c8e2" class="release-feature"> <p><strong>Vertex AI Model Garden</strong></p> <p>The Vertex AI Model Garden is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. The Model Garden is a platform that helps you discover, test, customize, and deploy Vertex AI and select OSS models. These models range from tunable to task-specific - all available on the Model Garden page in the Google Cloud console.</p> <ul> <li>To get started, see <a href="https://cloud.google.com/vertex-ai/docs/start/explore-models#create-a-vision-app">Explore AI models and APIs in Model Garden</a>.</li> </ul> </div> <h2 id="May_09_2023" data-text="May 09, 2023" tabindex="-1">May 09, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c0aa974c" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">G2 accelerator-optimized machine types</a> to serve predictions. Each G2 machine has a fixed number of NVIDIA L4 GPUs attached.</p> </div> <h2 id="May_04_2023" data-text="May 04, 2023" tabindex="-1">May 04, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="f2031b34" class="release-feature"> <p><strong>M108 release</strong></p> <p>The M108 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Miscellaneous software updates.</li> </ul> </div> <h2 id="April_14_2023" data-text="April 14, 2023" tabindex="-1">April 14, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="626779df" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now update some scaling and container logging configuration settings on a <a href="https://cloud.google.com/vertex-ai/docs/reference/rest/v1/projects.locations.endpoints#deployedmodel"><code translate="no" dir="ltr">DeployedModel</code></a> without undeploying and redeploying it to an endpoint.</p> <p>For more information, see <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#update_the_scaling_configuration">update the scaling configuration</a> and <a href="https://cloud.google.com/vertex-ai/docs/predictions/online-prediction-logging">container logging</a>.</p> </div> <h2 id="April_13_2023" data-text="April 13, 2023" tabindex="-1">April 13, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1da0c07e" class="release-announcement"> <p>The Timeseries Insights API is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). With the Timeseries Insights API, you can forecast and detect anomalies over billions of events in real time. For more information, see <a href="https://cloud.google.com/timeseries-insights/docs/overview">Timeseries Insights</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="2cf57325" class="release-feature"> <p><strong>M107 release</strong></p> <p>The M107 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Fixed a bug that displayed the wrong version of the JupyterLab user interface.</li> <li>Fixed a bug where a cron job for the diagnostic tool was added at every restart.</li> <li>Miscellaneous software updates.</li> </ul> </div> <h2 id="April_06_2023" data-text="April 06, 2023" tabindex="-1">April 06, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="a63a1402" class="release-feature"> <p><strong>M106 release</strong></p> <p>The M106 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Rolled back a previous change in which Jupyter dependencies were located in a separate Conda environment.</li> <li>Fixed a bug in which kernels used by notebooks did not contain the specified machine learning frameworks.</li> <li>Miscellaneous software updates.</li> </ul> </div> <h2 id="April_04_2023" data-text="April 04, 2023" tabindex="-1">April 04, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="195add05" class="release-feature"> <p>The Vertex AI Matching Engine service now offers <a href="https://cloud.google.com/products#product-launch-stages">Preview</a> support for deploying an index to a public endpoint. For information about how to get started, see <a href="https://cloud.google.com/vertex-ai/docs/matching-engine/match-eng-setup/match-eng-setup.md">Matching Engine Setup</a>.</p> </div> <div id="cb905fc3" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now view logs for Vertex AI <a href="https://cloud.google.com/vertex-ai/docs/predictions/overview#batch_predictions">Batch Prediction</a> jobs in <a href="https://cloud.google.com/logging/docs/overview">Cloud Logging</a>.</p> </div> <div id="bc6c7ff0" class="release-feature"> <p>Vertex AI Pipelines is now integrated with <a href="https://cloud.google.com/asset-inventory/docs/overview">Cloud Asset Inventory</a> service. You can use Cloud Asset Inventory to search, export, monitor, and analyze pipeline resources and metadata, and also view the resource history.</p> </div> <h2 id="April_03_2023" data-text="April 03, 2023" tabindex="-1">April 03, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2687860c" class="release-feature"> <p>The Vertex AI Model Registry now offers <a href="https://cloud.google.com/products#product-launch-stages">Preview</a> support for model copy between regions. For information about how to copy your model between regions, see <a href="https://cloud.google.com/vertex-ai/docs/model-registry/copy-model">Copy models in Model Registry</a>.</p> </div> <h2 id="March_31_2023" data-text="March 31, 2023" tabindex="-1">March 31, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="b8438109" class="release-feature"> <p><strong>M105 release</strong></p> <p>The M105 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li><p>The following user-managed notebooks images are now available with Python 3.10 on Debian 11:</p> <ul> <li>TensorFlow 2.11 CPU (<code translate="no" dir="ltr">tf-2-11-cpu-debian-11-py310</code>)</li> <li>TensorFlow 2.11 GPU with Cuda 11.3 (<code translate="no" dir="ltr">tf-2-11-cu113-notebooks-debian-11-py310</code>)</li> <li>PyTorch 1.13 with Cuda 11.3 (<code translate="no" dir="ltr">pytorch-1-13-cu113-notebooks-debian-11-py310</code>)</li> <li>Base CPU (<code translate="no" dir="ltr">common-cpu-notebooks-debian-11-py310</code>)</li> <li>Base GPU with Cuda 11.3 (<code translate="no" dir="ltr">common-cu113-notebooks-debian11-py310</code>)</li> </ul></li> <li><p>The following user-managed notebooks images are now available with Python 3.9 on Debian 11:</p> <ul> <li>TensorFlow 2.6 CPU (<code translate="no" dir="ltr">tf-2-6-cpu-notebooks-debian-11-py39</code>)</li> <li>TensorFlow 2.6 GPU with Cuda 11.3 (<code translate="no" dir="ltr">tf-2-6-cu113-notebooks-debian-11-py39</code>)</li> </ul></li> <li><p>Jupyter-related libraries have been moved to a different Conda environment, separate from the one containing machine learning frameworks and base software libraries.</p></li> </ul> </div> <h2 id="March_28_2023" data-text="March 28, 2023" tabindex="-1">March 28, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a76408b2" class="release-feature"> <p>Vertex AI Pipelines cost showback with billing labels is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). You can now use billing labels to review the cost of a pipeline run, along with the cost of individual resources generated from Google Cloud Pipeline Components in the pipeline run. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/understand-pipeline-cost-labels">Understand pipeline run costs</a>. </p> </div> <h2 id="March_27_2023" data-text="March 27, 2023" tabindex="-1">March 27, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="a2b095f0" class="release-feature"> <p><strong>M105 release</strong></p> <p>The M105 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed an issue wherein a runtime with idle shutdown enabled doesn't detect activity and shuts down.</li> <li>Fixed an issue wherein the runtime data disk runs out of space and prevents access.</li> <li>Fixed an issue wherein end user credentials are not preserved after shutdown.</li> <li>Changed Health Agent logging levels from <code translate="no" dir="ltr">DEBUG</code> to <code translate="no" dir="ltr">INFO</code>.</li> </ul> </div> <h2 id="March_21_2023" data-text="March 21, 2023" tabindex="-1">March 21, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="21b155c5" class="release-feature"> <p>Vertex AI supports running Explainable AI on certain types of BQML models when they are added to the Vertex AI Model Registry <a href="https://cloud.google.com/products#product-launch-stages">(GA)</a>. To learn more, see <a href="https://cloud.google.com/bigquery-ml/docs/vertex-xai">Explainable AI for BigQuery ML models</a>.</p> </div> <div id="4116c133" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p>The ability to delete feature values from an entity type is now generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). The following features are available:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/featurestore/delete-feature-values#delete_from_entities">Delete feature values from specified entities</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/featurestore/delete-feature-values#delete_from_features_timerange">Delete feature values from specified features within a time range</a></li> </ul> <p>Links to additional resources:</p> <ul> <li><a href="https://cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform_v1.services.featurestore_service.FeaturestoreServiceClient#google_cloud_aiplatform_v1_services_featurestore_service_FeaturestoreServiceClient_delete_feature_values">Learn more about the <code translate="no" dir="ltr">delete_feature_values</code> method in the Vertex AI SDK for Python.</a></li> <li><a href="https://github.com/googleapis/python-aiplatform/blob/main/samples/generated_samples/aiplatform_v1_generated_featurestore_service_delete_feature_values_sync.py">View the Python code sample on GitHub</a></li> <li><a href="https://github.com/googleapis/python-aiplatform/blob/main/samples/generated_samples/aiplatform_v1_generated_featurestore_service_delete_feature_values_async.py">View the Python (Async) code sample on GitHub</a></li> </ul> </div> <h2 id="March_20_2023" data-text="March 20, 2023" tabindex="-1">March 20, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d67ee0b1" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">N2, N2D, C2, and C2D machine types</a> to serve predictions.</p> </div> <h2 id="March_16_2023" data-text="March 16, 2023" tabindex="-1">March 16, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="43c26254" class="release-feature"> <p><strong>M104 release</strong></p> <p>The M104 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Fixed a regression in which <code translate="no" dir="ltr">jupyter-user</code> metadata was ignored.</li> <li>Enabled access to the Jupyter Gateway Client configuration by using the <code translate="no" dir="ltr">notebook-enable-gateway-client</code> and <code translate="no" dir="ltr">gateway-client-url</code> metadata tags.</li> <li>Added the following packages: <ul> <li>google-cloud-artifact-registry</li> <li>google-cloud-bigquery-storage</li> <li>google-cloud-language</li> <li>keyring</li> <li>keyrings.google-artifactregistry-auth</li> </ul></li> <li>Fixed a bug in which curl could not find the right SSL certificate path by default.</li> </ul> </div> <div id="5a35f834" class="release-changed"> <p>TensorFlow Enterprise 2.1 has reached the end of its support period. See <a href="https://cloud.google.com/tensorflow-enterprise/docs/overview#version_details">Version details</a>.</p> </div> <h2 id="March_03_2023" data-text="March 03, 2023" tabindex="-1">March 03, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="86015695" class="release-changed"> <p>Pre-built containers to perform custom training with <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">TensorFlow 2.11</a>, <a href="https://cloud.google.com/vertex-ai/docs/training/pre-built-containers#pytorch">PyTorch 1.12</a>, or <a href="https://cloud.google.com/vertex-ai/docs/training/pre-built-containers#pytorch">PyTorch 1.13</a> are now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="February_28_2023" data-text="February 28, 2023" tabindex="-1">February 28, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d60273ab" class="release-changed"> <p>A new <a href="https://cloud.google.com/vertex-ai/docs/training/overview">custom training overview</a> page is available. The new overview page covers the following topics:</p> <ul> <li>What is custom training?</li> <li>Benefits of custom training on Vertex AI.</li> <li>How custom training works.</li> <li>Custom training workflow.</li> </ul> </div> <h2 id="February_21_2023" data-text="February 21, 2023" tabindex="-1">February 21, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="f4739a9e" class="release-feature"> <p><strong>M104 update</strong></p> <p>This update of the M104 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a bug where local and remote kernels are not displayed. This happens when remote kernels are not accessible. </li> <li>Minor bug fixes and improvements.</li> </ul> </div> <h2 id="February_14_2023" data-text="February 14, 2023" tabindex="-1">February 14, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b8776bd8" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers#pytorch">Pre-built PyTorch containers</a> for serving predictions from PyTorch models is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <div id="c8f674b6" class="release-feature"> <p>Vertex AI Matching Engine now supports Private Service Connect in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. To learn how to set up a a Private Service Connect instance, see <a href="https://cloud.google.com/vertex-ai/docs/matching-engine/match-eng-setup/private-service-connect">Using Private Service Connect</a>.</p> </div> <h2 id="February_13_2023" data-text="February 13, 2023" tabindex="-1">February 13, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f310a3e2" class="release-feature"> <p>Support for resource-level IAM policies for Vertex AI <code translate="no" dir="ltr">featurestore</code> and <code translate="no" dir="ltr">entityType</code> resources is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/resource-policy">Control access to resources</a>.</p> </div> <h2 id="February_10_2023" data-text="February 10, 2023" tabindex="-1">February 10, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f85393e6" class="release-fixed"> <p>When performing distributed training, Vertex AI properly sets the primary replica in <code translate="no" dir="ltr">CLUSTER_SPEC</code> as <code translate="no" dir="ltr">workerpool0</code> instead of <code translate="no" dir="ltr">chief</code>. For details, see <a href="https://cloud.google.com/vertex-ai/docs/training/distributed-training#cluster-spec-format">Format CLUSTER_SPEC</a>.</p> </div> <h2 id="February_09_2023" data-text="February 09, 2023" tabindex="-1">February 09, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="5a6b4be6" class="release-feature"> <p><strong>M104 release</strong></p> <p>The M104 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Added a fix for a security vulnerability in single-user managed notebooks instances.</li> <li>Made enhancements to the network selection user experience in the managed notebooks executor.</li> <li>Minor bug fixes and improvements.</li> </ul> </div> <h2 id="February_06_2023" data-text="February 06, 2023" tabindex="-1">February 06, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4f7c1b04" class="release-feature"> <p>The Vertex AI Pipelines <strong>Template Gallery</strong> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. You can bootstrap your MLOps workflows with Google-authored pipeline and component templates. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/use-template-gallery">Use a prebuilt template from the Template Gallery</a>.</p> </div> <h2 id="January_30_2023" data-text="January 30, 2023" tabindex="-1">January 30, 2023</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="638618c2" class="release-feature"> <p><strong>M103 release</strong></p> <p>The M103 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>Fixed a bug in which a warning tells the user to run <code translate="no" dir="ltr">jupyter lab build</code> when creating a new instance.</li> <li>Upgraded PyTorch to 1.13.1.</li> <li>Minor bug fixes and improvements.</li> </ul> </div> <h2 id="January_26_2023" data-text="January 26, 2023" tabindex="-1">January 26, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="44da0cdd" class="release-feature"> <p><strong>Tabular Workflow for End-to-End AutoML</strong> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/e2e-automl">Tabular Workflow for End-to-End AutoML</a>. </p> </div> <h2 id="January_18_2023" data-text="January 18, 2023" tabindex="-1">January 18, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="0ec92f36" class="release-changed"> <p><strong>Vertex AI Explainability</strong></p> <p>When uploading TensorFlow 2 models, the <a href="https://cloud.google.com/vertex-ai/docs/reference/rest/v1/ExplanationSpec#ExplanationMetadata"><code translate="no" dir="ltr">ExplanationMetadata</code></a> field is now optional, making it easier to configure your model for explainability. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/explainable-ai/configuring-explanations-feature-based#import-model-example">Import a model with an <code translate="no" dir="ltr">explanationSpec</code> field</a>.</p> </div> <h2 id="January_11_2023" data-text="January 11, 2023" tabindex="-1">January 11, 2023</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="0727ddce" class="release-changed"> <p>Vertex AI Matching Engine is available in the following regions:</p> <ul> <li><code translate="no" dir="ltr">us-west2</code> – (Los Angeles)</li> <li><code translate="no" dir="ltr">us-west3</code> – (Salt Lake City)</li> <li><code translate="no" dir="ltr">northamerica-northeast1</code> – (Montréal)</li> <li><code translate="no" dir="ltr">northamerica-northeast2</code> – (Toronto)</li> <li><code translate="no" dir="ltr">europe-central2</code> – (Warsaw)</li> <li><code translate="no" dir="ltr">europe-west2</code> – (London)</li> <li><code translate="no" dir="ltr">europe-west3</code> – (Frankfurt)</li> <li><code translate="no" dir="ltr">europe-west6</code> – (Zurich)</li> <li><code translate="no" dir="ltr">asia-east1</code> – (Taiwan)</li> <li><code translate="no" dir="ltr">Asia-east2</code> – (Hong Kong)</li> <li><code translate="no" dir="ltr">me-west1</code> – (Tel aviv) </li> </ul> <p>To see all of the available locations for Matching Engine, see the <a href="https://cloud.google.com/vertex-ai/docs/general/locations">Vertex AI Locations page</a>.</p> </div> <h2 id="December_20_2022" data-text="December 20, 2022" tabindex="-1">December 20, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2aa165c6" class="release-feature"> <p><strong>Vertex AI TensorFlow Profiler</strong></p> <p>Vertex AI TensorFlow Profiler is generally available <a href="https://cloud.google.com/products#section-23">GA</a>. You can use TensorFlow Profiler to debug model training performance for your custom training jobs.</p> <p>For details, see <a href="https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-profiler">Profile model training performance using Profiler</a>.</p> </div> <div id="7d9958d5" class="release-feature"> <p><strong>Vertex AI Matching Engine</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/matching-engine/overview">Vertex AI Matching Engine</a> now offers <a href="https://cloud.google.com/products/#product-launch-stages">General Availability</a> support for updating your indices using Streaming Update, which is real-time indexing for the Approximate Nearest Neighbor (ANN) service.</p> </div> <div id="67f71304" class="release-feature"> <p>Vertex AI Feature Store <a href="https://cloud.google.com/vertex-ai/docs/featurestore/ingesting-stream">streaming ingestion</a> is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <div id="5ef5ff4e" class="release-changed"> <p>You can now override the default data retention limit of 4000 days for the online store and the offline store in Vertex AI Feature Store.</p> <ul> <li>You can set the data retention limit for the online store at the featurestore level.</li> <li>You can set the data retention limit for the offline store at the entity type level.</li> </ul> </div> <h2 id="December_15_2022" data-text="December 15, 2022" tabindex="-1">December 15, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="e5ec5d95" class="release-feature"> <p><strong>M102 release</strong></p> <p>The M102 release of Vertex AI Workbench user-managed notebooks includes the following:</p> <ul> <li>TensorFlow 2.11 is now available.</li> <li>PyTorch 1.13 is now available.</li> <li>Regular security patches and package upgrades.</li> </ul> </div> <h2 id="December_09_2022" data-text="December 09, 2022" tabindex="-1">December 09, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="53b45215" class="release-feature"> <p><strong>M101 release</strong></p> <p>The M101 release of Vertex AI Workbench includes the following:</p> <ul> <li>TensorFlow patch version upgrades: <ul> <li>From 2.8.3 to 2.8.4.</li> <li>From 2.9.2 to 2.9.3.</li> <li>From 2.10.0 to 2.10.1.</li> </ul></li> <li>TensorFlow 1.15 on Vertex AI Workbench is now deprecated.</li> <li>Added <code translate="no" dir="ltr">*.notebooks.cloud.google.com</code> as part of the domains required for users to access Notebooks API. Removed <code translate="no" dir="ltr">*.datalab.cloud.google.com</code>.</li> <li>Regular security patches and package upgrades.</li> </ul> </div> <h2 id="December_05_2022" data-text="December 05, 2022" tabindex="-1">December 05, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b455ce4a" class="release-feature"> <p>The Pipeline Templates feature is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). The <strong>Your Templates</strong> tab is supported by Artifact Registry and allows you to publish and curate pipeline and component templatess. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/pipelines/create-pipeline-template">Create, upload, and use a pipeline template</a>.</p> </div> <h2 id="November_30_2022" data-text="November 30, 2022" tabindex="-1">November 30, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c3f18e60" class="release-feature"> <p><strong>AutoML image model updates</strong></p> <p>AutoML image classification and object detection now support a higher-accuracy model type. This model is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> <p>For information about how to train a model using the higher accuracy model type, see <a href="https://cloud.google.com/vertex-ai/docs/tutorials/image-recognition-automl/training#2_begin_automl_model_training">Begin AutoML model training</a>.</p> <p>Batch prediction is currently not supported for this model type.</p> </div> <div id="f0e462d0" class="release-feature"> <p>Cloud Logging for Vertex AI Pipelines is now generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/logging">View pipeline job logs</a>.</p> </div> <h2 id="November_18_2022" data-text="November 18, 2022" tabindex="-1">November 18, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e895f309" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now perform some simple filtering and transformation on the batch input in your <a href="https://cloud.google.com/vertex-ai/docs/reference/rest/v1beta1/projects.locations.batchPredictionJobs"><code translate="no" dir="ltr">BatchPredictionJob</code></a> requests without having to write any code in the prediction container. This feature is in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/predictions/get-predictions#filter_and_transform_input_data_preview">Filter and transform input data</a>.</p> </div> <h2 id="November_17_2022" data-text="November 17, 2022" tabindex="-1">November 17, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f475b856" class="release-feature"> <p>The Vertex AI Pipelines email notification component is now generally available (<a href="https://cloud.google.com/products/#product-launch-stages">GA</a>). This component enables you to configure your pipeline to send up to three emails upon success or failure of a pipeline run. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/pipelines/email-notifications">Configure email notifications</a> and the <a href="https://cloud.google.com/vertex-ai/docs/pipelines/email-notification-component">Email notification component</a>.</p> </div> <h2 id="November_16_2022" data-text="November 16, 2022" tabindex="-1">November 16, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="45b7a156" class="release-feature"> <p>Vertex AI has added support for the following regions:</p> <ul> <li><code translate="no" dir="ltr">us-west3</code> (Salt Lake City)</li> <li><code translate="no" dir="ltr">europe-central2</code> (Warsaw)</li> <li><code translate="no" dir="ltr">asia-southeast2</code> (Jakarta)</li> <li><code translate="no" dir="ltr">me-west1</code> (Tel aviv) </li> </ul> <p>Some features of Vertex AI are not supported in these regions. Check feature availability for all regions on the <a href="https://cloud.google.com/vertex-ai/docs/general/locations">Vertex AI Locations page</a>.</p> </div> <h2 id="November_10_2022" data-text="November 10, 2022" tabindex="-1">November 10, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="82d2f00b" class="release-feature"> <p><strong>AutoML Image Classification Error Analysis</strong></p> <p>Error analysis allows you to examine error cases after training a model from within the model evaluation page. This feature is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> <p>For each image you can inspect similar images from the training set to help identify the following:</p> <ul> <li>Label inconsistencies between visually similar images</li> <li>Outliers if a test sample has no visually similar images in the training set</li> </ul> <p>After fixing any data issues, you can retrain the model to improve model performance.</p> </div> <h2 id="November_09_2022" data-text="November 09, 2022" tabindex="-1">November 09, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="03d2a537" class="release-feature"> <p>Feature Transform Engine is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/feature-engineering">Feature engineering</a>.</p> </div> <h2 id="November_08_2022" data-text="November 08, 2022" tabindex="-1">November 08, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="11c6375b" class="release-feature"> <p><strong>M100 release</strong></p> <p>The M100 release of Vertex AI Workbench includes the following:</p> <ul> <li>Fixed a bug that prevented an instance with a GPU from starting.</li> <li>Regular package updates.</li> <li>Miscellaneous bug and display fixes.</li> </ul> </div> <div id="664d5083" class="release-fixed"> <p>Fixed a server-side request forgery (SSRF) vulnerability. Previous versions of managed notebooks and user-managed notebooks instances still contain the vulnerability. It is recommended that you <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/migrate">migrate your data</a> to a new instance.</p> </div> <h2 id="November_04_2022" data-text="November 04, 2022" tabindex="-1">November 04, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4fa57df1" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">A2 machine types</a> to serve predictions.</p> </div> <div id="d8b0ebf1" class="release-feature"> <p><strong>Vertex ML Metadata</strong></p> <p>You can now <a href="https://cloud.google.com/vertex-ai/docs/ml-metadata/analyzing#filter_contexts_executions_and_artifacts_by_association_and_attribution">filter contexts, executions, and artifacts by association and attribution</a>.</p> </div> <div id="9daa31fa" class="release-feature"> <p>Custom training on Vertex AI now supports <a href="https://cloud.google.com/compute/docs/gpus#a100-80gb">NVIDIA A100 80GB GPUs</a> on <code translate="no" dir="ltr">a2-ultragpu-1g/2g/4g/8g</code> machines. For details, see <a href="https://cloud.google.com/vertex-ai/docs/training/configure-compute">Configure compute resources for custom training</a>.</p> </div> <h2 id="November_03_2022" data-text="November 03, 2022" tabindex="-1">November 03, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2c661485" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/predictions/custom-prediction-routines">Custom prediction routines (CPR)</a> are now <a href="https://cloud.google.com/products/#product-launch-stages">Generally Available</a>. CPR lets you easily build custom containers for prediction with pre/post processing support.</p> </div> <h2 id="October_27_2022" data-text="October 27, 2022" tabindex="-1">October 27, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="21678857" class="release-feature"> <p><strong>Vertex AI Prediction</strong></p> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/predictions/configure-compute#machine-types">E2 machine types</a> to serve predictions.</p> </div> <h2 id="October_25_2022" data-text="October 25, 2022" tabindex="-1">October 25, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="1faaba15" class="release-breaking"> <p>The <code translate="no" dir="ltr">v1beta1</code> version of the Notebooks API is scheduled for removal no earlier than January 16, 2023. After this date, you must use Notebooks API <code translate="no" dir="ltr">v1</code> to manage Vertex AI Workbench resources.</p> </div> <h2 id="October_18_2022" data-text="October 18, 2022" tabindex="-1">October 18, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="8f241498" class="release-feature"> <p><strong>M98 release</strong></p> <p>The M98 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Upgraded Go from 1.16.5 to 1.19.2.</li> <li>Upgraded R from 4.1 to 4.2.</li> <li>Upgraded JupyterLab from 3.2 to 3.4.</li> <li>Miscellaneous bug and display fixes.</li> <li>Added a fix for the BigQuery SQL editor to run queries correctly in non-US locations.</li> <li>Regular package updates.</li> </ul> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/versions">Learn more about managed notebooks versions</a>.</p> </div> <h2 id="October_12_2022" data-text="October 12, 2022" tabindex="-1">October 12, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6cd9e979" class="release-feature"> <p>Tabular Workflow for TabNet Training is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/tabnet">Tabular Workflows for TabNet Training</a>. </p> <p>Tabular Workflow for Wide & Deep Training is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/wide-and-deep">Tabular Workflow for Wide & Deep Training</a>. </p> </div> <h2 id="October_11_2022" data-text="October 11, 2022" tabindex="-1">October 11, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d18a8202" class="release-feature"> <p>Vertex AI Feature Store <a href="https://cloud.google.com/vertex-ai/docs/featurestore/ingesting-stream">streaming ingestion</a> is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <h2 id="October_10_2022" data-text="October 10, 2022" tabindex="-1">October 10, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c6362d12" class="release-feature"> <p><b>The Vertex AI Model Registry</b> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). <a href="http://cloud.google.com/vertex-ai/docs/model-registry/introduction">Vertex AI Model Registry</a> is a searchable repository where you can manage the lifecycle of your ML models. From the Vertex AI Model Registry, you can better organize your models, train new versions, and deploy directly to endpoints.</p> </div> <div id="c86f8413" class="release-feature"> <p>The <b>Vertex AI Model Registry and BigQuery ML</B> integration is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). With this integration, <a href="https://cloud.google.com/bigquery-ml/docs/managing-models-vertex">BigQuery ML</a> models can be managed alongside other ML models in Vertex AI to easily version, evaluate, and deploy for prediction.</p> </div> <h2 id="October_06_2022" data-text="October 06, 2022" tabindex="-1">October 06, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2d56fa62" class="release-feature"> <p><strong>Incrementally train an AutoML model</strong></p> <p>You can now incrementally train an AutoML image classification or object detection model by selecting a previously trained model. This feature is in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/tutorials/image-recognition-automl/training">Train an AutoML image classification model</a>.</p> </div> <h2 id="October_05_2022" data-text="October 05, 2022" tabindex="-1">October 05, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="cb5cb681" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p>The ability to <a href="https://cloud.google.com/vertex-ai/docs/featurestore/delete-feature-values">delete feature values</a> from an entity type is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. The following features are available:</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/featurestore/delete-feature-values#delete_from_entities">Delete feature values from specified entities</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/featurestore/delete-feature-values#delete_from_features_timerange">Delete feature values from specified features within a time range</a></li> </ul> <p>Links to additional resources:</p> <ul> <li><a href="https://cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform_v1beta1.services.featurestore_service.FeaturestoreServiceClient#google_cloud_aiplatform_v1beta1_services_featurestore_service_FeaturestoreServiceClient_delete_feature_values">Learn more about the <code translate="no" dir="ltr">delete_feature_values</code> method in the Vertex AI SDK for Python.</a></li> <li><a href="https://github.com/googleapis/python-aiplatform/blob/9a506eeae6be3fa33b262dc7782d0bb4f10bf765/samples/generated_samples/aiplatform_v1beta1_generated_featurestore_service_delete_feature_values_sync.py">View the Python code sample on GitHub</a></li> <li><a href="https://github.com/googleapis/python-aiplatform/blob/9a506eeae6be3fa33b262dc7782d0bb4f10bf765/samples/generated_samples/aiplatform_v1beta1_generated_featurestore_service_delete_feature_values_async.py">View the Python (Async) code sample on GitHub</a></li> </ul> </div> <h2 id="October_04_2022" data-text="October 04, 2022" tabindex="-1">October 04, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e8da45a1" class="release-feature"> <p><strong>Vertex AI model evaluation</strong> is now available in Preview. <a href="https://cloud.google.com/vertex-ai/docs/evaluation/introduction">Model evaluation</a> provides model evaluation metrics, such as precision and recall, to help you determine the performance of your models. </p> </div> <h2 id="September_26_2022" data-text="September 26, 2022" tabindex="-1">September 26, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="474900ae" class="release-feature"> <p><strong>Vertex AI Model Monitoring</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/model-monitoring">Vertex AI Model Monitoring</a> now offers <a href="https://cloud.google.com/products#product-launch-stages">Preview</a> support for batch prediction jobs. For more details, see <a href="https://cloud.google.com/vertex-ai/docs/model-monitoring/model-monitoring-batch-predictions">Vertex AI Model Monitoring for batch predictions</a>. </p> </div> <div id="0b95c0fd" class="release-feature"> <p><strong>Vertex AI Feature Store</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/featurestore/monitoring#feature">Feature value monitoring</a> is now generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="September_22_2022" data-text="September 22, 2022" tabindex="-1">September 22, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="637451d7" class="release-feature"> <p><strong>Vertex AI Matching Engine</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/matching-engine/overview">Vertex AI Matching Engine</a> now offers <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a> support for updating your indices using Streaming Update, which is real-time indexing for the Approximate Nearest Neighbor (ANN) service.</p> </div> <h2 id="September_20_2022" data-text="September 20, 2022" tabindex="-1">September 20, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bb7e788e" class="release-feature"> <p>The option to configure pipeline run caching (<a href="https://cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform.PipelineJob"><code translate="no" dir="ltr">enable_caching</code></a>) is now available in the Cloud console.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="d2a2fd5f" class="release-feature"> <p><strong>M96 release</strong></p> <p>The M96 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a problem where users were not able to save large Notebooks.</li> <li>Fixed a display issue when using JupyterLab's simple interface.</li> <li>Improved timeout behavior switch hardware operations.</li> <li>Improved error messaging when a service account cannot access the Runtime.</li> <li>Security fixes.</li> <li>Regular package refreshment and bug fixes. </li> </ul> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/versions">Learn more about managed notebooks versions</a>.</p> </div> <div id="b8608cb4" class="release-fixed"> <p>Fixed a server-side request forgery (SSRF) vulnerability. Previous versions of managed notebooks and user-managed notebooks instances still contain the vulnerability. It is recommended that you <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/migrate">migrate your data</a> to a new instance.</p> </div> <h2 id="September_14_2022" data-text="September 14, 2022" tabindex="-1">September 14, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="be2e7998" class="release-feature"> <p>You can now limit the number of concurrent or parallel task runs in a pipeline run using <code translate="no" dir="ltr">dsl.ParallelFor</code>. For more information, see the <a href="https://kubeflow-pipelines.readthedocs.io/en/sdk-2.0.1/source/dsl.html?h=dsl.parallelfor">Kubeflow Pipelines SDK Documentation</a>.</p> </div> <div id="1136a301" class="release-changed"> <p>The performance of the <code translate="no" dir="ltr">ListPipelineJobs</code> API has been improved via a new <a href="https://cloud.google.com/vertex-ai/docs/reference/rest/v1/projects.locations.pipelineJobs/list#query-parameters"><code translate="no" dir="ltr">readMask</code></a> that lets you filter out large fields. To leverage this in the Python SDK, use the new <a href="https://cloud.google.com/python/docs/reference/aiplatform/latest/google.cloud.aiplatform.PipelineJob#google_cloud_aiplatform_PipelineJob_list"><code translate="no" dir="ltr">enable_simple_view</code></a>. </p> </div> <h2 id="August_17_2022" data-text="August 17, 2022" tabindex="-1">August 17, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="a53d7caa" class="release-feature"> <p><strong>M95 release</strong></p> <p>The M95 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Fixed a bug where users were regularly getting a 502 error when trying to access JupyterLab. </li> <li>Fixed a bug where opening an instance in Single User mode slowed the start of an instance. </li> <li>Fixed a bug where a managed notebooks instance was not starting after adding a GPU. </li> <li>Fixed bugs on the Serverless Spark form input. </li> <li>Improved the ActivityLog refresh after Serverless Spark creation. </li> <li>Fixed a bug related to the display of materialized views in BigQuery. </li> <li>Refreshed the JupyterLab interface with an improved Google-specific theme. </li> <li>Fixed a bug related to viewing Cloud Storage buckets and folders with large numbers of objects. </li> <li>Regular package refreshment and bug fixes. </li> </ul> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/versions">Learn more about managed notebooks versions</a>.</p> </div> <h2 id="August_12_2022" data-text="August 12, 2022" tabindex="-1">August 12, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e2b22d39" class="release-feature"> <p><strong>Vertex Explainable AI</strong></p> <p><a href="https://cloud.google.com/vertex-ai/docs/explainable-ai/overview">Vertex Explainable AI</a> now offers <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a> support for example-based explanations. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/explainable-ai/configuring-explanations-example-based">Configure example-based explanations for custom training</a>.</p> </div> <h2 id="August_01_2022" data-text="August 01, 2022" tabindex="-1">August 01, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6a1909b4" class="release-feature"> <p><strong>TensorFlow Profiler integration</strong>: Debug model training performance for your custom training jobs. For details, see <a href="https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-profiler">Profile model training performance using Profiler</a>.</p> </div> <h2 id="July_29_2022" data-text="July 29, 2022" tabindex="-1">July 29, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1a0ead1d" class="release-feature"> <p>Vertex AI now offers <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a> support for <a href="https://cloud.google.com/vertex-ai/docs/predictions/custom-prediction-routines">Custom prediction routines (CPR)</a>. CPR lets you easily build custom containers for prediction with pre/post processing support.</p> </div> <h2 id="July_18_2022" data-text="July 18, 2022" tabindex="-1">July 18, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4de3d8b7" class="release-feature"> <p>NFS support for custom training is GA. For details, see <a href="https://cloud.google.com/vertex-ai/docs/training/train-nfs-share">Mount an NFS share for custom training</a>.</p> </div> <h2 id="July_14_2022" data-text="July 14, 2022" tabindex="-1">July 14, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="48cb7423" class="release-feature"> <p>The <strong>Pipeline Templates</strong> feature is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/pipelines/create-pipeline-template">Create, upload, and use a pipeline template</a>.</p> <p>The features supported by pipeline templates include the following:</p> <ul> <li>Create a template registry using Artifact Registry (AR).</li> <li>Compile and publish a pipeline template.</li> <li>Create a pipeline run using the template and filter the runs.</li> <li>Manage (create, update, or delete) the pipeline template resources.</li> </ul> </div> <h2 id="July_12_2022" data-text="July 12, 2022" tabindex="-1">July 12, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2cb0b0a3" class="release-changed"> <p>You can now use a pre-built container to perform custom training with <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">TensorFlow 2.9</a></p> </div> <h2 id="July_11_2022" data-text="July 11, 2022" tabindex="-1">July 11, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="537b5440" class="release-feature"> <p>Vertex AI Pipelines now lets you configure task-level retries. You can set the number of times a task is retried before it fails. For more information about this option, see the <a href="https://kubeflow-pipelines.readthedocs.io/en/stable/source/dsl.html#kfp.dsl.PipelineTask.set_retry">Kubeflow Pipelines SDK Documentation</a>.</p> </div> <h2 id="July_06_2022" data-text="July 06, 2022" tabindex="-1">July 06, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="62995945" class="release-feature"> <p><strong>Tabular Workflows</strong> is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/overview">Tabular Workflows on Vertex AI</a>. </p> <p><strong>End-to-End AutoML</strong> workflow is available in <a href="https://cloud.google.com/products#product-launch-stages">Public Preview</a>. For documentation, refer to <a href="https://cloud.google.com/vertex-ai/docs/tabular-data/tabular-workflows/e2e-automl">End-to-End AutoML</a>. </p> </div> <h2 id="June_30_2022" data-text="June 30, 2022" tabindex="-1">June 30, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="72c67eb7" class="release-feature"> <p>Feature: <a href="https://cloud.google.com/vertex-ai/docs/experiments/intro-vertex-ai-experiments">Vertex AI Experiments</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Vertex AI Experiments helps users track and compare multiple experiment runs and analyze key model metrics.</p> <p>Features supported by Experiments include:</p> <ul> <li>Vary and track parameters and metrics.</li> <li>Compare parameters, metrics, and artifacts between pipeline runs.</li> <li>Track steps and artifacts to capture the lineage of experiments.</li> <li>Compare vertex pipelines against Notebook experiments.</li> </ul> </div> <h2 id="June_28_2022" data-text="June 28, 2022" tabindex="-1">June 28, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1ca37b44" class="release-feature"> <p><strong>Vertex AI Forecasting</strong> is available in <a href="https://cloud.google.com/products/#product-launch-stages">GA</a>. The following features are available: </p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting-arima/overview">Forecasting with ARIMA+</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting-explanations">Explainable AI for Forecasting</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/hierarchical">Hierarchical Forecasting</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/train-model">Model training with AutoML or Seq2Seq+</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/train-model">Forecasting window configuration during model training</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/tabular-data/forecasting/train-model">Holiday effect modeling during model training</a></li> </ul> </div> <h2 id="June_17_2022" data-text="June 17, 2022" tabindex="-1">June 17, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4404a956" class="release-feature"> <p>Support for IAM <a href="https://cloud.google.com/vertex-ai/docs/featurestore/resource-policy">resource-level policies</a> for Vertex AI featurestore and entityType resources is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <h2 id="May_27_2022" data-text="May 27, 2022" tabindex="-1">May 27, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="58588b97" class="release-feature"> <p><strong>M93 release</strong></p> <p>The M93 release of Vertex AI Workbench managed notebooks includes the following:</p> </div> <div id="58215963" class="release-fixed"> <ul> <li>Fixed a bug that prevented kernels from shutting down properly in Vertex AI Workbench managed notebooks.</li> </ul> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/versions">Learn more about managed notebooks versions</a>.</p> </div> <h2 id="May_24_2022" data-text="May 24, 2022" tabindex="-1">May 24, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fab1e69c" class="release-feature"> <p>You can now <a href="https://cloud.google.com/vertex-ai/docs/pipelines/configure-failure-policy">configure the failure policy</a> for a pipeline run.</p> </div> <h2 id="May_18_2022" data-text="May 18, 2022" tabindex="-1">May 18, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ff16d33d" class="release-feature"> <p>The ability to configure Vertex AI private endpoints is now general available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Vertex AI private endpoints provide a low-latency, secure connection to the Vertex AI online prediction service. You can configure Vertex AI private endpoints by using VPC Network Peering. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints">Use private endpoints for online prediction</a>.</p> </div> <h2 id="May_12_2022" data-text="May 12, 2022" tabindex="-1">May 12, 2022</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="6e5dd844" class="release-feature"> <p><strong>M91 release</strong></p> <p>The M91 release of Vertex AI Workbench managed notebooks includes the following:</p> <ul> <li>Log streaming to the consumer project via Logs Viewer is now supported.</li> <li>Added the <code translate="no" dir="ltr">net-tools</code> package.</li> <li>Regular package refreshments and bug fixes.</li> </ul> </div> <div id="22661bde" class="release-fixed"> <ul> <li>Fixed an issue that caused Spark server networking errors when using Dataproc Serverless Spark and VPC Peering.</li> </ul> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/versions">Learn more about managed notebooks versions</a>.</p> </div> <h2 id="April_26_2022" data-text="April 26, 2022" tabindex="-1">April 26, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="99dadc5b" class="release-feature"> <p>You can now train your custom models using <a href="https://cloud.google.com/vertex-ai/docs/training/training-with-tpu-vm">Cloud TPU Architecture (TPU VMs)</a>.</p> </div> <h2 id="April_21_2022" data-text="April 21, 2022" tabindex="-1">April 21, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="39b4a7ee" class="release-feature"> <p>You can now use a pre-built container to perform custom training with <a href="https://cloud.google.com/vertex-ai/docs/training/pre-built-containers#pytorch">PyTorch 1.11</a>.</p> </div> <h2 id="April_06_2022" data-text="April 06, 2022" tabindex="-1">April 06, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="7c75a397" class="release-feature"> <p><a href="http://cloud.google.com/vertex-ai/docs/model-registry/introduction">Vertex AI Model Registry</a> is available in <a href="https://cloud.google.com/products/#product-launch-stages">Preview</a>. Vertex AI Model Registry is a searchable repository where you can manage the lifecycle of your ML models. From the Vertex AI Model Registry, you can better organize your models, train new versions, and deploy directly to endpoints.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="068dbec4" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench">Vertex AI Workbench</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Vertex AI Workbench is a single notebook surface for all your data science needs that lets you access BigQuery data and Cloud Storage from within JupyterLab, execute notebook code in Vertex AI custom training and Spark, use custom containers, manage costs with idle timeout, and secure your instances with VPC Service Controls and customer managed encryption keys (CMEK). </p> <p>Features supported include:</p> <ul> <li>Google-managed instances and the latest GPU support</li> <li><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/idle-shutdown">Idle shutdown</a> for managed notebooks instances</li> <li><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/custom-container">Custom containers</a></li> <li>End-user and service account authentication</li> <li>Native plug-ins for <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/bigquery">BigQuery</a> and <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/cloud-storage">Cloud Storage</a></li> <li>In-notebook Spark connect to Dataproc clusters</li> <li>Jobs support via the <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/executor">managed notebooks executor</a> on Vertex AI custom training and Spark</li> <li>One-click deploy for NGC containers </li> <li>VPC Service Controls</li> <li><a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/cmek">Customer managed encryption keys (CMEK)</a></li> </ul> </div> <div id="f909e446" class="release-feature"> <p>The Vertex AI Workbench managed notebooks <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/executor">executor</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Use the executor to run notebook files on a schedule or as a one-time execution. You can use parameters in your execution to make specific changes to each run. For example, you might specify a different dataset to use, change the learning rate on your model, or change the version of the model. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/managed/executor">Run notebook files with the executor</a>.</p> </div> <h2 id="March_07_2022" data-text="March 07, 2022" tabindex="-1">March 07, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9ef80143" class="release-feature"> <p>Vertex AI Feature Store <a href="https://cloud.google.com/vertex-ai/docs/featurestore/managing-featurestores#scaling_options">online store autoscaling</a> is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. The online store nodes automatically scale to balance performance and cost with different traffic patterns. The offline store already scales automatically.</p> </div> <div id="26e8c7f7" class="release-feature"> <p>You can now mount Network File System (NFS) shares to access remote files when you run a custom training job. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/training/train-nfs-share">Mount an NFS share for custom training</a>.</p> <p>This feature is in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <div id="ce721f7b" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/pipelines/components-introduction">Google Cloud Pipeline Components SDK</a> v1.0 is now generally available.</p> </div> <h2 id="February_16_2022" data-text="February 16, 2022" tabindex="-1">February 16, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6d58f433" class="release-feature"> <p>You can now use a pre-built container to perform custom training with <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">TensorFlow 2.8</a>.</p> </div> <h2 id="February_10_2022" data-text="February 10, 2022" tabindex="-1">February 10, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="12a54baf" class="release-changed"> <p>For Vertex AI featurestore resources, the online store is optional. You can set the number of online nodes to <code translate="no" dir="ltr">0</code>. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/featurestore/managing-featurestores">Manage featurestores</a>.</p> </div> <h2 id="January_04_2022" data-text="January 04, 2022" tabindex="-1">January 04, 2022</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fa1d6650" class="release-feature"> <p>You can now use a pre-built container to perform custom training with <a href="https://cloud.google.com/vertex-ai/docs/training/pre-built-containers#pytorch">PyTorch 1.10</a>.</p> </div> <h2 id="December_23_2021" data-text="December 23, 2021" tabindex="-1">December 23, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="6c25a055" class="release-changed"> <p>There are now three Vertex AI release note feeds. Add any of the following to your <a href="https://wikipedia.org/wiki/Comparison_of_feed_aggregators">feed reader</a>:</p> <ul> <li>For both Vertex AI and Vertex AI Workbench: <code translate="no" dir="ltr">https://cloud.google.com/feeds/vertex-ai-product-group-release-notes.xml</code></li> <li>For Vertex AI only: <code translate="no" dir="ltr">https://cloud.google.com/feeds/vertex-ai-release-notes.xml</code></li> <li>For Vertex AI Workbench only: <code translate="no" dir="ltr">https://cloud.google.com/feeds/aiplatformnotebooks-release-notes.xml</code></li> </ul> </div> <h2 id="December_02_2021" data-text="December 02, 2021" tabindex="-1">December 02, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="185f1de5" class="release-feature"> <p>You can now use a pre-built container to perform custom training with <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">TensorFlow 2.7</a>.</p> </div> <h2 id="December_01_2021" data-text="December 01, 2021" tabindex="-1">December 01, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="dbb340ef" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-overview">Vertex AI TensorBoard</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="November_19_2021" data-text="November 19, 2021" tabindex="-1">November 19, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bff5fc2b" class="release-feature"> <p>The <a href="https://cloud.google.com/vertex-ai/docs/training/create-custom-job">autopackaging</a> feature of the <code translate="no" dir="ltr">gcloud ai custom-jobs create</code> command is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). Autopackaging lets you use a single command to run code on your local computer as a custom training job in Vertex AI.</p> </div> <div id="55879493" class="release-feature"> <p>The <code translate="no" dir="ltr">gcloud ai customs-jobs local-run</code> command is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). You can use this command to <a href="https://cloud.google.com/vertex-ai/docs/training/containerize-run-code-local">containerize and run training code locally</a>.</p> </div> <h2 id="November_09_2021" data-text="November 09, 2021" tabindex="-1">November 09, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9daeeab0" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/pipelines/introduction">Vertex AI Pipelines</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>). </p> </div> <h2 id="November_02_2021" data-text="November 02, 2021" tabindex="-1">November 02, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a4d99dc0" class="release-feature"> <p>Using <a href="https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell">interactive shells</a> to inspect custom training jobs is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> <p>You can use these interactive shells with <a href="https://cloud.google.com/vertex-ai/docs/general/vpc-service-controls">VPC Service Controls</a>.</p> </div> <h2 id="October_25_2021" data-text="October 25, 2021" tabindex="-1">October 25, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="67554a79" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/ml-metadata">Vertex ML Metadata</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="October_11_2021" data-text="October 11, 2021" tabindex="-1">October 11, 2021</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="855294a0" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench">Vertex AI Workbench</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex AI Workbench is a notebook-based development environment for the entire data science workflow.</p> </div> <div id="648ba049" class="release-announcement"> <p>The Notebooks product and all existing Notebooks instances are now part of Vertex AI Workbench as <a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/introduction">user-managed notebooks</a>.</p> </div> <h2 id="October_05_2021" data-text="October 05, 2021" tabindex="-1">October 05, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f688c7da" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/featurestore/overview">Vertex Feature Store</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="September_24_2021" data-text="September 24, 2021" tabindex="-1">September 24, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f2d6137b" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/matching-engine/overview">Vertex Matching Engine</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="September_21_2021" data-text="September 21, 2021" tabindex="-1">September 21, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="2eede73c" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/vizier/overview">Vertex AI Vizier</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="September_15_2021" data-text="September 15, 2021" tabindex="-1">September 15, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="9a62e052" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/explainable-ai/overview">Vertex Explainable AI</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <h2 id="September_13_2021" data-text="September 13, 2021" tabindex="-1">September 13, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="10df828e" class="release-feature"> <p>You can use a <a href="https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers#tensorflow">pre-built container to serve predictions from TensorFlow 2.6 models</a>.</p> </div> <h2 id="September_10_2021" data-text="September 10, 2021" tabindex="-1">September 10, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="4fca7661" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/model-monitoring/overview">Vertex Model Monitoring</a> is generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>).</p> </div> <div id="115cee1e" class="release-feature"> <p>When you perform custom training, you can <a href="https://cloud.google.com/vertex-ai/docs/training/code-requirements#fuse">access Cloud Storage buckets by reading and writing to the local filesystem</a>. This feature, based on <a href="https://cloud.google.com/storage/docs/gcs-fuse">Cloud Storage Fuse</a>, is available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="5bae8b23" class="release-changed"> <p>Due to a recent change, the <code translate="no" dir="ltr">iam.serviceAccounts.actAs</code> permission on the specified service account for the notebook instance is required for users to continue to have access to their notebook instances. The Google internal Inverting Proxy server that provides access to notebook instances now verifies that this permission is present before allowing users access to the JupyterLab URL. The JupyterLab URL this update covers is: <code translate="no" dir="ltr">*.notebooks.googleusercontent.com</code> This update only applies to notebook instances in <strong>Single User</strong> mode and verifies that the assigned single user is authorized to execute code inside the notebook instance. Notebook instances running in Service Account or Project Editor mode already perform this verification via the Inverting Proxy server.</p> </div> <h2 id="August_30_2021" data-text="August 30, 2021" tabindex="-1">August 30, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="f1a00c8b" class="release-feature"> <p>You can now use a pre-built container to perform custom training with <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">TensorFlow 2.6</a> and <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#pytorch">PyTorch 1.9</a>.</p> </div> <h2 id="August_24_2021" data-text="August 24, 2021" tabindex="-1">August 24, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="68b388af" class="release-feature"> <p>The following tools for creating embeddings to use with <a href="https://cloud.google.com/vertex-ai/docs/matching-engine/overview">Vertex Matching Engine</a> are available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>:</p> <ul> <li>the Two Tower built-in algorithm </li> <li>the Swivel pipeline template</li> </ul> </div> <h2 id="August_02_2021" data-text="August 02, 2021" tabindex="-1">August 02, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b2659aba" class="release-changed"> <p>Vertex Pipelines is available in the following regions:</p> <ul> <li><code translate="no" dir="ltr">us-east1</code> (South Carolina)</li> <li><code translate="no" dir="ltr">europe-west2</code> (London)</li> <li><code translate="no" dir="ltr">asia-southeast1</code> (Singapore)</li> </ul> <p>See all the <a href="https://cloud.google.com/vertex-ai/docs/general/locations">locations where Vertex Pipelines is available</a>.</p> </div> <h2 id="July_28_2021" data-text="July 28, 2021" tabindex="-1">July 28, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="01b1641a" class="release-feature"> <p>You can <a href="https://cloud.google.com/vertex-ai/docs/training/distributed-training#reducing_training_time_with_reduction_server">use the Reduction Server algorithm</a> (<a href="https://cloud.google.com/products#product-launch-stages">Preview</a>) to increase throughput and reduce latency during distributed custom training.</p> </div> <h2 id="July_27_2021" data-text="July 27, 2021" tabindex="-1">July 27, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="53228346" class="release-feature"> <p>The following features are generally available (<a href="https://cloud.google.com/products#product-launch-stages">GA</a>):</p> <ul> <li><a href="https://cloud.google.com/vertex-ai/docs/general/access-transparency">Access Transparency for Vertex AI</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/general/custom-service-account">Using a custom service account for custom training and prediction</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/general/vpc-service-controls">Using VPC Service Controls with Vertex AI</a></li> <li><a href="https://cloud.google.com/vertex-ai/docs/general/vpc-peering">Setting up VPC Network Peering with Vertex AI</a> and <a href="https://cloud.google.com/vertex-ai/docs/training/using-private-ip">using private IP for custom training</a> (Using <a href="https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints">private IP for prediction</a> and <a href="https://cloud.google.com/vertex-ai/docs/matching-engine/using-matching-engine">vector matching with Matching Engine</a> remains in preview.)</li> </ul> </div> <h2 id="July_26_2021" data-text="July 26, 2021" tabindex="-1">July 26, 2021</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1 & v1</div> <div id="d7362a55" class="release-changed"> <p>If using proxy single-user mode, Notebooks API now verifies if the specified user (<code translate="no" dir="ltr">proxy-user-mail</code>) has Service Account permissions on the Service Account. This check is performed during instance creation and registration.</p> </div> <h2 id="July_20_2021" data-text="July 20, 2021" tabindex="-1">July 20, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="17533752" class="release-feature"> <p>Private endpoints for online prediction are now available in preview. After you <a href="https://cloud.google.com/vertex-ai/docs/general/vpc-peering">set up VPC Network Peering with Vertex AI</a>, you can <a href="https://cloud.google.com/vertex-ai/docs/predictions/using-private-endpoints">create private endpoints for low-latency online prediction</a> within your private network.</p> <p>Additionally, the documentation for VPC Network Peering with custom training has moved. The general instructions for setting up VPC Network Peering with Vertex AI are available at the original link, https://cloud.google.com/vertex-ai/docs/general/vpc-peering. The documentation for custom training is now available here: <a href="https://cloud.google.com/vertex-ai/docs/training/using-private-ip">Using private IP with custom training</a>.</p> </div> <h2 id="July_19_2021" data-text="July 19, 2021" tabindex="-1">July 19, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="ce07b558" class="release-feature"> <p>You can now <a href="https://cloud.google.com/vertex-ai/docs/training/monitor-debug-interactive-shell">use an interactive shell</a> to inspect your custom training container while it runs. The interactive shell can be helpful for monitoring and debugging training.</p> <p>This feature is available in preview.</p> </div> <h2 id="July_14_2021" data-text="July 14, 2021" tabindex="-1">July 14, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="bcd2e86b" class="release-feature"> <p>You can now <a href="https://cloud.google.com/vertex-ai/docs/training/create-custom-job#create">use the <code translate="no" dir="ltr">gcloud beta ai custom-jobs create</code> command</a> to build a Docker image based on local training code, push the image to Container Registry, and create a <code translate="no" dir="ltr">CustomJob</code> resource.</p> </div> <h2 id="July_08_2021" data-text="July 08, 2021" tabindex="-1">July 08, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1ca13c1d" class="release-feature"> <p>You can now <a href="https://cloud.google.com/vertex-ai/docs/training/containerize-run-code-local">containerize and run your training code locally</a> by using the new <code translate="no" dir="ltr">gcloud beta ai custom-jobs local-run</code> command. This feature is available in preview.</p> </div> <h2 id="June_25_2021" data-text="June 25, 2021" tabindex="-1">June 25, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="0238725e" class="release-feature"> <p>You can now use <a href="https://cloud.google.com/vertex-ai/docs/training/configure-compute#specifying_gpus">NVIDIA A100 GPUs</a> and several <a href="https://cloud.google.com/vertex-ai/docs/training/configure-compute#machine-types">accelerator-optimized (A2) machine types</a> for training. You must use A100 GPUs and A2 machine types together. Learn about <a href="https://cloud.google.com/vertex-ai/pricing">their pricing</a>.</p> </div> <h2 id="June_18_2021" data-text="June 18, 2021" tabindex="-1">June 18, 2021</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="533a8203" class="release-feature"> <p>Support for Compute Reservations. Notebooks API allows the use of Compute Reservations during instance creation.</p> </div> <h2 id="June_11_2021" data-text="June 11, 2021" tabindex="-1">June 11, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d5e4442c" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers#tensorflow">pre-built container to serve predictions from TensorFlow 2.5 models</a>.</p> </div> <div id="b3d9cd70" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/vertex-ai/docs/predictions/pre-built-containers#xgboost">pre-built container to serve predictions from XGBoost 1.4 models</a>.</p> </div> <h2 id="May_18_2021" data-text="May 18, 2021" tabindex="-1">May 18, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="c1989e79" class="release-announcement"> <p>AI Platform (Unified) is now <a href="https://cloud.google.com/vertex-ai/">Vertex AI</a>.</p> </div> <div id="507a64bd" class="release-feature"> <p>Vertex AI has added support for custom model training, custom model batch prediction, custom model online prediction, and a limited number of other services in the following <a href="https://cloud.google.com/vertex-ai/docs/general/locations">regions</a>:</p> <ul> <li>us-west1</li> <li>us-east1</li> <li>us-east4</li> <li>northamerica-northeast1</li> <li>europe-west2</li> <li>europe-west1</li> <li>asia-southeast1</li> <li>asia-northeast1</li> <li>australia-southeast1 </li> <li>asia-northeast3</li> </ul> </div> <div id="ac413769" class="release-feature"> <p>Vertex AI now supports <a href="https://cloud.google.com/vertex-ai/docs/start/automl-model-types#tabular">forecasting with time series data</a> for AutoML tabular models, in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. You can use forecasting to predict a series of numeric values that extend into the future.</p> </div> <div id="192d4dab" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/pipelines/introduction">Vertex Pipelines</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex Pipelines helps you to automate, monitor, and govern your ML systems by orchestrating your ML workflow.</p> </div> <div id="1e31562d" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/model-monitoring/overview">Vertex Model Monitoring</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex Model Monitoring enables you to monitor model quality over time.</p> </div> <div id="6b6577e2" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/featurestore/overview">Vertex Feature Store</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex Feature Store provides a centralized repository for organizing, storing, and serving ML features.</p> </div> <div id="d20a39a1" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/ml-metadata/introduction">Vertex ML Metadata</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex ML Metadata lets you record the metadata and artifacts produced by your ML system so you can analyze the performance of your ML system.</p> </div> <div id="6eab6864" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/matching-engine/overview">Vertex Matching Engine</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex Matching Engine enables vector similarity search.</p> </div> <div id="21afc98e" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/experiments/tensorboard-overview">Vertex TensorBoard</a> is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>. Vertex TensorBoard enables you to track, visualize, and compare ML experiments.</p> </div> <h2 id="May_03_2021" data-text="May 03, 2021" tabindex="-1">May 03, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="967ac060" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers#tensorflow">pre-built container to serve predictions from TensorFlow 2.4 models</a>.</p> </div> <div id="51593a18" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers#scikit-learn">pre-built container to serve predictions from scikit-learn 0.24 models</a>.</p> </div> <div id="1b85e911" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers#xgboost">pre-built container to serve predictions from XGBoost 1.3 models</a>.</p> </div> <h2 id="April_27_2021" data-text="April 27, 2021" tabindex="-1">April 27, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="955c536d" class="release-feature"> <p><a href="https://cloud.google.com/ai-platform-unified/docs/vizier/overview">AI Platform Vizier</a> is now available in preview. Vizier is a feature of AI Platform (Unified) that you can use to perform black-box optimization. You can use Vizier to tune hyperparameters or optimize any evaluable system.</p> </div> <h2 id="April_15_2021" data-text="April 15, 2021" tabindex="-1">April 15, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="7465e577" class="release-feature"> <p>The Python client library for AI Platform (Unified) is now called the AI Platform (Unified) SDK. With the release of version 0.7 (<a href="https://github.com/googleapis/google-cloud-python/blob/master/README.rst#beta-support">Preview</a>), the AI Platform (Unified) SDK provides two levels of support. The high-level <code translate="no" dir="ltr">aiplatform</code> library is designed to simplify common data science workflows by using wrapper classes and opinionated defaults. The lower-level <code translate="no" dir="ltr">aiplatform.gapic</code> library remains available for those times when you need more flexibility or control. <a href="https://cloud.google.com/ai-platform-unified/docs/start/client-libraries">Learn more</a>.</p> </div> <h2 id="March_31_2021" data-text="March 31, 2021" tabindex="-1">March 31, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="b72a78b5" class="release-announcement"> <p>AI Platform (Unified) is now available in General Availability (GA).</p> </div> <div id="c1492fba" class="release-changed"> <p>AI Platform (Unified) has added support for the following regions for custom model training, as well as batch and online prediction for custom-trained models:</p> <ul> <li>us-west1 (Oregon)</li> <li>us-east1 (South Carolina)</li> <li>us-east4 (N. Virginia)</li> <li>northamerica-northeast1 (Montreal)</li> <li>europe-west2 (London)</li> <li>europe-west1 (Belgium)</li> <li>asia-southeast1 (Singapore)</li> <li>asia-northeast1 (Tokyo)</li> <li>australia-southeast1 (Sydney)</li> <li>asia-northeast3 (Seoul)</li> </ul> </div> <h2 id="March_26_2021" data-text="March 26, 2021" tabindex="-1">March 26, 2021</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="c5c47767" class="release-feature"> <p>Cross Project Service Account is supported for user-managed notebooks.</p> </div> <h2 id="March_15_2021" data-text="March 15, 2021" tabindex="-1">March 15, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="24a6e920" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#pytorch">pre-built container to perform custom training with PyTorch 1.7</a>.</p> </div> <h2 id="March_04_2021" data-text="March 04, 2021" tabindex="-1">March 04, 2021</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="152a0780" class="release-changed"> <p>New Notebooks instances add labels for VM image (<code translate="no" dir="ltr">goog-caip-notebook</code>) and volume (<code translate="no" dir="ltr">goog-caip-notebook-volume</code>).</p> </div> <h2 id="March_02_2021" data-text="March 02, 2021" tabindex="-1">March 02, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d87aee00" class="release-changed"> <p><strong>CMEK compliance using the client libraries</strong></p> <p>You can now use the client libraries to create resources with a customer-managed encryption key (CMEK).</p> <p>For more information on creating a resource with an encryption key using the client libraries, see <a href="https://cloud.google.com/ai-platform-unified/docs/general/cmek#create_resources_with_the_kms_key">Using customer-managed encryption keys (CMEK)</a>.</p> </div> <h2 id="March_01_2021" data-text="March 01, 2021" tabindex="-1">March 01, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="67c51316" class="release-changed"> <p>The <a href="https://cloud.google.com/ai-platform-unified/docs/start/client-libraries">client library</a> for Java now includes enhancements to improve usage of training and prediction features. The client library includes additional types and utility functions for sending training requests, sending prediction requests, and reading prediction results.</p> <p>To use these enhancements, you must install the latest version of the <a href="https://cloud.google.com/ai-platform-unified/docs/start/client-libraries">client library</a>.</p> </div> <h2 id="February_25_2021" data-text="February 25, 2021" tabindex="-1">February 25, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="1e46d5b4" class="release-feature"> <p>AI Platform (Unified) now supports Access Transparency in beta. Google Cloud organizations with certain support packages can use this feature. Learn more about <a href="https://cloud.google.com/ai-platform-unified/docs/general/access-transparency">using Access Transparency with AI Platform (Unified)</a>.</p> </div> <div id="a8246bb8" class="release-changed"> <p>The <a href="https://cloud.google.com/ai-platform-unified/docs/start/client-libraries">client libraries</a> for Node.js and Python now include enhancements to improve usage of training and prediction features. These client libraries include additional types and utility functions for sending training requests, sending prediction requests, and reading prediction results.</p> <p>To use these enhancements, you must install the latest version of the <a href="https://cloud.google.com/ai-platform-unified/docs/start/client-libraries">client libraries</a>.</p> </div> <div id="2b74c051" class="release-changed"> <p>The <code translate="no" dir="ltr">predict</code> and <code translate="no" dir="ltr">explain</code> method calls no longer require the use of a different service endpoint (for example, <code translate="no" dir="ltr">https://us-central1-prediction-aiplatform.googleapis.com</code>). These methods are now available on the same endpoint as all other methods.</p> </div> <div id="8f1ebb50" class="release-feature"> <p>In addition to Docker images hosted on <a href="https://cloud.google.com/container-registry/docs/">Container Registry</a>, you can now use Docker images hosted on <a href="https://cloud.google.com/artifact-registry/docs/">Artifact Registry</a> and <a href="https://hub.docker.com/">Docker Hub</a> for <a href="https://cloud.google.com/ai-platform-unified/docs/training/containers-overview">custom container training on AI Platform</a>.</p> </div> <div id="7d96380c" class="release-changed"> <p>The Docker images for <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers">pre-built training containers</a> and <a href="https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers">pre-built prediction containers</a> are now available on Artifact Registry.</p> </div> <div id="58fc1e86" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">pre-built container to perform custom training with TensorFlow 2.4</a>.</p> </div> <div id="e445042a" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers#tensorflow">pre-built container to serve predictions from TensorFlow 2.3 models</a>.</p> </div> <div id="58051367" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/predictions/pre-built-containers#xgboost">pre-built container to serve predictions from XGBoost 1.2 models</a>.</p> </div> <h2 id="February_01_2021" data-text="February 01, 2021" tabindex="-1">February 01, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e278499d" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#pytorch">pre-built container to perform custom training with PyTorch 1.6</a>.</p> </div> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="bb232dab" class="release-feature"> <p>Notebooks <a href="https://registry.terraform.io/providers/hashicorp/google/latest/docs/resources/notebooks_instance">Terraform Module</a> supports Notebooks API v1</p> </div> <h2 id="January_23_2021" data-text="January 23, 2021" tabindex="-1">January 23, 2021</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="350c10a0" class="release-announcement"> <p>VPC-SC for Notebooks (now known as user-managed notebooks) is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>.</p> </div> <div id="85d54f73" class="release-feature"> <p>Notebooks API supports <a href="https://cloud.google.com/shielded-vm">Shielded VM configuration</a>.</p> </div> <h2 id="January_19_2021" data-text="January 19, 2021" tabindex="-1">January 19, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="5334f14c" class="release-feature"> <p><strong>Preview</strong>: Select AI Platform (Unified) resources can now be configured to use <a href="https://cloud.google.com/ai-platform-unified/docs/general/cmek">Customer-managed encryption keys (CMEK)</a>.</p> <p>Currently you can only create resources with a CMEK key in the UI; this functionality is not currently available using the client libraries. </p> </div> <h2 id="January_11_2021" data-text="January 11, 2021" tabindex="-1">January 11, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="e0b2ffd9" class="release-breaking"> <p>The default boot disk type for virtual machine instances used for custom training has changed from <code translate="no" dir="ltr">pd-standard</code> to <code translate="no" dir="ltr">pd-ssd</code>. Learn more about <a href="https://cloud.google.com/ai-platform-unified/docs/training/configure-compute#specifying_boot_disk_options">disk types for custom training</a> and read about <a href="https://cloud.google.com/ai-platform-unified/pricing#training">pricing for different disk types</a>.</p> <p>If you previously used the default disk type for custom training and want to continue training with the same disk type, make sure to explicitly specify the <code translate="no" dir="ltr">pd-standard</code> boot disk type when you perform custom training.</p> </div> <h2 id="January_06_2021" data-text="January 06, 2021" tabindex="-1">January 06, 2021</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="d75e1494" class="release-feature"> <p>You can now use a <a href="https://cloud.google.com/ai-platform-unified/docs/training/pre-built-containers#tensorflow">pre-built container to perform custom training with TensorFlow 2.3</a>.</p> </div> <h2 id="December_17_2020" data-text="December 17, 2020" tabindex="-1">December 17, 2020</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="a1026cf9" class="release-feature"> <p>AI Platform (Unified) now stores and processes your data only in the region you specify for most features. <a href="https://cloud.google.com/ai-platform-unified/docs/general/locations">Learn more</a>.</p> </div> <h2 id="November_16_2020" data-text="November 16, 2020" tabindex="-1">November 16, 2020</h2> <strong class="release-note-product-title">Vertex AI</strong> <div class="release-note-product-version-title">v1</div> <div id="fa85e1ca" class="release-feature"> <p><strong>Preview release</strong></p> <p>AI Platform (Unified) is now available in <a href="https://cloud.google.com/products#product-launch-stages">Preview</a>.</p> <p>For more information, see the <a href="https://cloud.google.com/ai-platform-unified/docs">product documentation</a>.</p> </div> <h2 id="September_21_2020" data-text="September 21, 2020" tabindex="-1">September 21, 2020</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="7fa55965" class="release-feature"> <p>AI Platform Notebooks (now known as user-managed notebooks) API is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>. The API now includes an <a href="https://cloud.google.com/ai-platform/notebooks/docs/reference/rest/v1/projects.locations.instances/isUpgradeable">isUpgradable</a> endpoint and adds manual and auto-upgrade functionality to notebooks instances created using the API.</p> </div> <div id="54e3ee78" class="release-feature"> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/audit-logging">Cloud Audit Logging for AI Platform Notebooks</a> (now known as user-managed notebooks) is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>.</p> <p><a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/iam">Granular IAM permissions for AI Platform Notebooks</a> (now known as user-managed notebooks) is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>.</p> <p>AI Platform Notebooks now supports <a href="https://cloud.google.com/compute/docs/machine-types#e2_machine_types">E2 machine types</a>.</p> <p>The following new regions have been added:</p> <ul> <li><code translate="no" dir="ltr">europe-west2</code> (London, UK) </li> <li><code translate="no" dir="ltr">europe-west3</code> (Frankfurt, Germany)</li> <li><code translate="no" dir="ltr">europe-west6</code> (Zürich, Switzerland)</li> </ul> </div> <h2 id="March_31_2020" data-text="March 31, 2020" tabindex="-1">March 31, 2020</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="2f63ad5e" class="release-feature"> <p>AI Platform Notebooks (now known as user-managed notebooks) is now <a href="https://cloud.google.com/products#product-launch-stages">Generally Available</a>. Some integrations with and specific features of AI Platform Notebooks are still in beta, such as <a href="https://cloud.google.com/ai-platform/notebooks/docs/service-perimeter">Virtual Private Cloud Service Controls</a>, <a href="https://cloud.google.com/ai-platform/notebooks/docs/iam">Identity and Access Management (IAM)</a> roles, and <a href="https://cloud.google.com/ai-platform/notebooks/docs/reference/rest">AI Platform Notebooks API</a>.</p> </div> <h2 id="February_04_2020" data-text="February 04, 2020" tabindex="-1">February 04, 2020</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="48cb4de5" class="release-feature"> <p><a href="https://cloud.google.com/vpc-service-controls/docs/overview">VPC Service Controls</a> now supports AI Platform Notebooks. Learn <a href="https://cloud.google.com/ai-platform/notebooks/docs/service-perimeter">how to use a notebook instance within a service perimeter</a>. This functionality is in <a href="https://cloud.google.com/products/#product-launch-stages">beta</a>.</p> </div> <h2 id="February_03_2020" data-text="February 03, 2020" tabindex="-1">February 03, 2020</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="13cc7983" class="release-feature"> <p>AI Platform Notebooks now supports Access Transparency. Access Transparency provides you with logs of actions that Google staff have taken when accessing your data. To learn more about Access Transparency, see the <a href="https://cloud.google.com/logging/docs/audit/access-transparency-overview#overview">Overview of Access Transparency</a>.</p> </div> <h2 id="September_12_2019" data-text="September 12, 2019" tabindex="-1">September 12, 2019</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1</div> <div id="2db9cc9d" class="release-feature"> <p>You can now use customer-managed encryption keys (CMEK) to protect data on the boot disks of your AI Platform Notebooks (now known as user-managed notebooks) VM instances. CMEK in AI Platform Notebooks is generally available. For more information, see <a href="https://cloud.google.com/vertex-ai/docs/workbench/user-managed/cmek">Using customer-managed encryption keys (CMEK)</a>.</p> </div> <h2 id="September_09_2019" data-text="September 09, 2019" tabindex="-1">September 09, 2019</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="7ffc5f06" class="release-feature"> <p>AI Platform Notebooks now provides more ways for you to customize your network settings, encrypt your notebook content, and grant access to your notebook instance. These options are available when you <a href="https://cloud.google.com/ai-platform/notebooks/docs/create-new#create-with-options">create a notebook</a>.</p> </div> <div id="08117351" class="release-feature"> <p>Now you can implement AI Platform Notebooks using custom containers. Use a <a href="https://cloud.google.com/ai-platform/deep-learning-containers/docs/choosing-container#choose_a_container_image_type">Deep Learning Containers image</a> or <a href="https://cloud.google.com/ai-platform/deep-learning-containers/docs/derivative-container">create a derivative container</a> of your own, then <a href="https://cloud.google.com/ai-platform/notebooks/docs/custom-container">create a new notebook instance using your custom container</a>.</p> </div> <h2 id="July_12_2019" data-text="July 12, 2019" tabindex="-1">July 12, 2019</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="2e1febb3" class="release-changed"> <p>R upgraded to version 3.6.</p> <p>R Notebooks are no longer dependent on a Conda environment.</p> </div> <h2 id="June_03_2019" data-text="June 03, 2019" tabindex="-1">June 03, 2019</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="2a9eec80" class="release-feature"> <p>You can now create AI Platform Notebooks instances with R and core R packages installed. Learn <a href="https://cloud.google.com/ai-platform/notebooks/docs/dependencies">how to install R dependencies</a>, and read guides for <a href="https://cloud.google.com/ai-platform/notebooks/docs/use-r-bigquery">using R with BigQuery in AI Platform Notebooks</a> and <a href="https://cloud.google.com/ai-platform/notebooks/docs/r-python-same-notebook">using R and Python in the same notebook</a>.</p> </div> <h2 id="March_01_2019" data-text="March 01, 2019" tabindex="-1">March 01, 2019</h2> <strong class="release-note-product-title">Vertex AI Workbench</strong> <div class="release-note-product-version-title">v1beta1</div> <div id="2f13d849" class="release-feature"> <p><a href="https://cloud.google.com/ai-platform/notebooks/docs/overview">AI Platform Notebooks</a> is now available in beta. AI Platform Notebooks enables you to create and manage virtual machine (VM) instances that are pre-packaged with <a href="https://jupyterlab.readthedocs.io/en/stable/getting_started/overview.html">JupyterLab</a> and a suite of deep learning software.</p> <p>Visit the <a href="https://cloud.google.com/ai-platform/notebooks/docs/overview">AI Platform Notebooks overview</a> and the <a href="https://cloud.google.com/ai-platform/notebooks/docs/create-new">guide to creating a new notebook instance</a> to learn more.</p> </div> </section> </section> <devsite-hats-survey class="nocontent" hats-id="mwETRvWii0eU5NUYprb0Y9z5GVbc" listnr-id="83405"></devsite-hats-survey> </div> <devsite-thumb-rating position="footer"> </devsite-thumb-rating> <devsite-feedback position="footer" project-name="Vertex AI" product-id="5164350" bucket="Documentation" context="" version="t-devsite-webserver-20250211-r00-rc00.466928320959134316" data-label="Send Feedback Button" track-type="feedback" track-name="sendFeedbackLink" track-metadata-position="footer" class="nocontent" project-feedback-url="https://issuetracker.google.com/issues/new?component=1130925" project-icon="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/favicons/onecloud/super_cloud.png" project-support-url="https://cloud.google.com/support-hub" > <button> Send feedback </button> </devsite-feedback> <div class="devsite-floating-action-buttons"> </div> </article> <devsite-content-footer class="nocontent"> <p>Except as otherwise noted, the content of this page is licensed under the <a href="https://creativecommons.org/licenses/by/4.0/">Creative Commons Attribution 4.0 License</a>, and code samples are licensed under the <a href="https://www.apache.org/licenses/LICENSE-2.0">Apache 2.0 License</a>. For details, see the <a href="https://developers.google.com/site-policies">Google Developers Site Policies</a>. Java is a registered trademark of Oracle and/or its affiliates.</p> <p>Last updated 2025-02-14 UTC.</p> </devsite-content-footer> <devsite-notification > </devsite-notification> <div class="devsite-content-data"> <template class="devsite-thumb-rating-feedback"> <devsite-feedback position="thumb-rating" project-name="Vertex AI" product-id="5164350" bucket="Documentation" context="" version="t-devsite-webserver-20250211-r00-rc00.466928320959134316" data-label="Send Feedback Button" track-type="feedback" track-name="sendFeedbackLink" track-metadata-position="thumb-rating" class="nocontent" project-feedback-url="https://issuetracker.google.com/issues/new?component=1130925" project-icon="https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/favicons/onecloud/super_cloud.png" project-support-url="https://cloud.google.com/support-hub" > <button> Need to tell us more? </button> </devsite-feedback> </template> <template class="devsite-content-data-template"> [[["Easy to understand","easyToUnderstand","thumb-up"],["Solved my problem","solvedMyProblem","thumb-up"],["Other","otherUp","thumb-up"]],[["Hard to understand","hardToUnderstand","thumb-down"],["Incorrect information or sample code","incorrectInformationOrSampleCode","thumb-down"],["Missing the information/samples I need","missingTheInformationSamplesINeed","thumb-down"],["Other","otherDown","thumb-down"]],["Last updated 2025-02-14 UTC."],[],[]] </template> </div> </devsite-content> </main> <devsite-footer-promos class="devsite-footer"> </devsite-footer-promos> <devsite-footer-linkboxes class="devsite-footer"> <nav class="devsite-footer-linkboxes nocontent" aria-label="Footer links"> <ul class="devsite-footer-linkboxes-list"> <li class="devsite-footer-linkbox "> <h3 class="devsite-footer-linkbox-heading no-link">Why Google</h3> <ul class="devsite-footer-linkbox-list"> <li class="devsite-footer-linkbox-item"> <a href="/why-google-cloud/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 1)" track-metadata-child_headline="why google"track-metadata-position="footer"track-type="footer link"track-name="choosing google cloud"track-metadata-eventDetail="cloud.google.com/why-google-cloud/"track-metadata-module="footer"> Choosing Google Cloud </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/trust-center/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 2)" track-metadata-module="footer"track-metadata-child_headline="why google"track-name="trust and security"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/security/"track-type="footer link"> Trust and security </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/modern-infrastructure/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 3)" track-type="footer link"track-name="modern infrastructure cloud"track-metadata-eventDetail="cloud.google.com/solutions/modern-infrastructure/"track-metadata-child_headline="why google"track-metadata-position="footer"track-metadata-module="footer"> Modern Infrastructure Cloud </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/multicloud/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 4)" track-metadata-position="footer"track-metadata-child_headline="why google"track-metadata-module="footer"track-metadata-eventDetail="cloud.google.com/multicloud/"track-name="multicloud"track-type="footer link"> Multicloud </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/infrastructure/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 5)" track-type="footer link"track-metadata-eventDetail="cloud.google.com/infrastructure/"track-name="global infrastructure"track-metadata-module="footer"track-metadata-position="footer"track-metadata-child_headline="why google"> Global infrastructure </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/customers/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 6)" track-type="footer link"track-metadata-module="footer"track-name="customers and case studies"track-metadata-child_headline="why google"track-metadata-eventDetail="cloud.google.com/customers/"track-metadata-position="footer"> Customers and case studies </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/analyst-reports/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 7)" track-metadata-position="footer"track-metadata-child_headline="why google"track-type="footer link"track-metadata-module="footer"track-name="analyst reports"track-metadata-eventDetail="cloud.google.com/analyst-reports/"> Analyst reports </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/whitepapers/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 8)" track-type="footer link"track-metadata-module="footer"track-name="whitepapers"track-metadata-eventDetail="cloud.google.com/whitepapers/"track-metadata-child_headline="why google"track-metadata-position="footer"> Whitepapers </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//cloud.google.com/blog/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 9)" track-metadata-module="footer"track-name="blog"track-type="footer link"track-metadata-eventDetail="cloud.google.com/blog/"track-metadata-child_headline="engage"track-metadata-position="footer"> Blog </a> </li> </ul> </li> <li class="devsite-footer-linkbox "> <h3 class="devsite-footer-linkbox-heading no-link">Products and pricing</h3> <ul class="devsite-footer-linkbox-list"> <li class="devsite-footer-linkbox-item"> <a href="/pricing/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 1)" track-metadata-eventDetail="cloud.google.com/pricing/"track-name="google cloud pricing"track-metadata-position="footer"track-type="footer link"track-metadata-module="footer"track-metadata-child_headline="products and pricing"> Google Cloud pricing </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//workspace.google.com/pricing.html" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 2)" track-metadata-eventDetail="workspace.google.com/pricing.html"track-type="footer link"track-metadata-position="footer"track-name="google workspace pricing"track-metadata-module="footer"track-metadata-child_headline="products and pricing"target="_blank"> Google Workspace pricing </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/products/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 3)" track-name="see all products"track-metadata-eventDetail="cloud.google.com/products/"track-type="footer link"track-metadata-child_headline="products and pricing"track-metadata-position="footer"track-metadata-module="footer"> See all products </a> </li> </ul> </li> <li class="devsite-footer-linkbox "> <h3 class="devsite-footer-linkbox-heading no-link">Solutions</h3> <ul class="devsite-footer-linkbox-list"> <li class="devsite-footer-linkbox-item"> <a href="/solutions/infrastructure-modernization/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 1)" track-type="footer link"track-name="infrastructure modernization"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/solutions/infrastructure-modernization/"track-metadata-child_headline="solutions"track-metadata-module="footer"> Infrastructure modernization </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/databases/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 2)" track-metadata-position="footer"track-type="footer link"track-metadata-child_headline="solutions"track-metadata-module="footer"track-metadata-eventDetail="cloud.google.com/solutions/databases"track-name="databases"> Databases </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/application-modernization/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 3)" track-metadata-child_headline="solutions"track-name="application development"track-metadata-module="footer"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/solutions/application-modernization/"track-type="footer link"> Application modernization </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/smart-analytics/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 4)" track-metadata-child_headline="solutions"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/solutions/smart-analytics/"track-name="smart analytics"track-metadata-module="footer"track-type="footer link"> Smart analytics </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/ai/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 5)" track-metadata-child_headline="solutions"track-name="artificial intelligence"track-metadata-position="footer"track-type="footer link"track-metadata-eventDetail="cloud.google.com/solutions/ai/"track-metadata-module="footer"> Artificial Intelligence </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/security/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 6)" track-metadata-position="footer"track-type="footer link"track-metadata-module="footer"track-metadata-eventDetail="cloud.google.com/solutions/security/"track-metadata-child_headline="solutions"track-name="security"> Security </a> </li> <li class="devsite-footer-linkbox-item"> <a href="https://workspace.google.com/enterprise/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 7)" track-metadata-eventDetail="workspace.google.com/enterprise/"track-metadata-child_headline="solutions"track-type="footer link"track-name="productivity and work transformation"track-metadata-module="footer"target="_blank"track-metadata-position="footer"> Productivity & work transformation </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/#industry-solutions" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 8)" track-metadata-eventDetail="cloud.google.com/solutions/#industry-solutions"track-name="industry solutions"track-metadata-module="footer"track-type="footer link"track-metadata-position="footer"track-metadata-child_headline="solutions"> Industry solutions </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/devops/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 9)" track-metadata-child_headline="solutions"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/solutions/devops/"track-name="devops solutions"track-metadata-module="footer"track-type="footer link"> DevOps solutions </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/#section-14" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 10)" track-metadata-child_headline="solutions"track-metadata-module="footer"track-name="small business solutions"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/solutions/#section-14"track-type="footer link"> Small business solutions </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/solutions/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 11)" track-metadata-module="footer"track-type="footer link"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/solutions/"track-metadata-child_headline="solutions"track-name="see all solutions"> See all solutions </a> </li> </ul> </li> <li class="devsite-footer-linkbox "> <h3 class="devsite-footer-linkbox-heading no-link">Resources</h3> <ul class="devsite-footer-linkbox-list"> <li class="devsite-footer-linkbox-item"> <a href="/affiliate-program/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 1)" track-metadata-eventDetail="cloud.google.com/affiliate-program/"track-name="google cloud affiliate program"track-metadata-child_headline="resources"track-metadata-module="footer"track-type="footer link"track-metadata-position="footer"> Google Cloud Affiliate Program </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/docs/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 2)" track-metadata-eventDetail="cloud.google.com/docs/"track-name="google cloud documentation"track-metadata-module="footer"track-type="footer link"track-metadata-child_headline="resources"track-metadata-position="footer"> Google Cloud documentation </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/docs/get-started/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 3)" track-metadata-child_headline="resources"track-metadata-module="footer"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/docs/get-started/"track-name="google cloud quickstarts"track-type="footer link"> Google Cloud quickstarts </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/marketplace/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 4)" track-metadata-position="footer"track-metadata-module="footer"track-metadata-eventDetail="cloud.google.com/marketplace/"track-metadata-child_headline="resources"track-type="footer link"track-name="google cloud marketplace"> Google Cloud Marketplace </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/discover/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 5)" track-metadata-module="footer"track-type="footer link"track-metadata-child_headline="resources"track-name="learn about cloud computing"track-metadata-position="footer"track-metadata-eventDetail="learn/"> Learn about cloud computing </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/support-hub/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 6)" track-name="support"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/support-hub/"track-metadata-module="footer"track-type="footer link"track-metadata-child_headline="resources"> Support </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/docs/samples" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 7)" track-metadata-module="footer"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/docs/samples"track-type="footer link"track-metadata-child_headline="resources"track-name="code samples"> Code samples </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/architecture/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 8)" track-name="cloud architecture center"track-type="footer link"track-metadata-module="footer"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/architecture/"track-metadata-child_headline="resources"> Cloud Architecture Center </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/learn/training/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 9)" track-metadata-module="footer"track-type="footer link"track-metadata-child_headline="resources"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/training/"track-name="training"> Training </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/learn/certification/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 10)" track-metadata-child_headline="resources"track-metadata-position="footer"track-metadata-module="footer"track-name="certifications"track-metadata-eventDetail="cloud.google.com/certification"track-type="footer link"> Certifications </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//developers.google.com" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 11)" track-name="google developers"track-type="footer link"target="_blank"track-metadata-position="footer"track-metadata-eventDetail="developers.google.com"track-metadata-child_headline="resources"track-metadata-module="footer"> Google for Developers </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/startup/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 12)" track-metadata-child_headline="resources"track-metadata-position="footer"track-name="google cloud for startups"track-type="footer link"track-metadata-eventDetail="cloud.google.com/startup/"track-metadata-module="footer"> Google Cloud for Startups </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//status.cloud.google.com" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 13)" track-name="system status"target="_blank"track-type="footer link"track-metadata-eventDetail="status.cloud.google.com"track-metadata-child_headline="resources"track-metadata-position="footer"track-metadata-module="footer"> System status </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/release-notes" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 14)" track-metadata-child_headline="resources"track-metadata-module="footer"track-name="release notes"track-type="footer link"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/release-notes/"> Release Notes </a> </li> </ul> </li> <li class="devsite-footer-linkbox "> <h3 class="devsite-footer-linkbox-heading no-link">Engage</h3> <ul class="devsite-footer-linkbox-list"> <li class="devsite-footer-linkbox-item"> <a href="/contact/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 1)" track-metadata-module="footer"track-type="footer link"track-metadata-child_headline="engage"track-name="contact sales"track-metadata-position="footer"track-metadata-eventDetail="cloud.google.com/contact/"> Contact sales </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//cloud.google.com/find-a-partner" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 2)" track-name="find a partner"target="_blank"track-metadata-position="footer"track-metadata-module="footer"track-type="footer link"track-metadata-child_headline="engage"track-metadata-eventDetail="cloud.google.com/find-a-partner"> Find a Partner </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/partners/become-a-partner/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 3)" track-name="become a partner"track-metadata-position="footer"track-type="footer link"track-metadata-eventDetail="cloud.google.com/partners/become-a-partner/"track-metadata-child_headline="engage"track-metadata-module="footer"> Become a Partner </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/events/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 4)" track-metadata-module="footer"track-metadata-position="footer"track-metadata-eventDetail="cloud.withgoogle.com/events"track-metadata-child_headline="engage"track-type="footer link"track-name="events"> Events </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/podcasts/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 5)" track-name="podcasts"target="_blank"track-metadata-position="footer"track-metadata-module="footer"rel="noopener"track-metadata-eventDetail="cloud.google.com/podcasts/"track-metadata-child_headline="engage"track-type="footer link"> Podcasts </a> </li> <li class="devsite-footer-linkbox-item"> <a href="/developers/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 6)" track-metadata-module="footer"track-metadata-child_headline="engage"track-name="developer center"track-metadata-position="footer"track-type="footer link"track-metadata-eventDetail="cloud.google.com/developers/"> Developer Center </a> </li> <li class="devsite-footer-linkbox-item"> <a href="https://www.googlecloudpresscorner.com/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 7)" track-metadata-position="footer"track-metadata-eventDetail="www.googlecloudpresscorner.com"track-type="footer link"track-metadata-module="footer"rel="noopener"track-metadata-child_headline="engage"track-name="press corner"target="_blank"> Press Corner </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//www.youtube.com/googlecloud" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 8)" track-type="footer link"track-metadata-child_headline="engage"track-name="google cloud on youtube"rel="noopener"track-metadata-position="footer"track-metadata-module="footer"track-metadata-eventDetail="www.youtube.com/googlecloud"target="_blank"> Google Cloud on YouTube </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//www.youtube.com/googlecloudplatform" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 9)" track-metadata-position="footer"track-metadata-module="footer"target="_blank"track-metadata-eventDetail="www.youtube.com/googlecloudplatform"rel="noopener"track-type="footer link"track-metadata-child_headline="engage"track-name="google cloud tech on youtube"> Google Cloud Tech on YouTube </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//x.com/googlecloud" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 10)" track-metadata-eventDetail="x.com/googlecloud"track-metadata-child_headline="engage"target="_blank"rel="noopener"track-name="follow on x"track-metadata-module="footer"track-type="footer link"track-metadata-position="footer"> Follow on X </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//userresearch.google.com/?reserved=1&utm_source=website&Q_Language=en&utm_medium=own_srch&utm_campaign=CloudWebFooter&utm_term=0&utm_content=0&productTag=clou&campaignDate=jul19&pType=devel&referral_code=jk212693" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 11)" target="_blank"track-type="footer link"track-metadata-module="footer"track-metadata-eventDetail="userresearch.google.com/?reserved=1&utm_source=website&Q_Language=en&utm_medium=own_srch&utm_campaign=CloudWebFooter&utm_term=0&utm_content=0&productTag=clou&campaignDate=jul19&pType=devel&referral_code=jk212693"track-metadata-child_headline="engage"track-metadata-position="footer"track-name="join user research"> Join User Research </a> </li> <li class="devsite-footer-linkbox-item"> <a href="//careers.google.com/cloud" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 12)" track-metadata-position="footer"track-metadata-module="footer"track-metadata-eventDetail="careers.google.com/cloud"track-type="footer link"target="_blank"track-metadata-child_headline="engage"track-name="we are hiring join google cloud"> We're hiring. Join Google Cloud! </a> </li> <li class="devsite-footer-linkbox-item"> <a href="https://www.googlecloudcommunity.com/" class="devsite-footer-linkbox-link gc-analytics-event" data-category="Site-Wide Custom Events" data-label="Footer Link (index 13)" rel="noopener"target="_blank"track-metadata-position="footer"track-metadata-module="footer"track-metadata-child_headline="engage"track-name="google cloud community"track-metadata-eventDetail="www.googlecloudcommunity.com"track-type="footer link"> Google Cloud Community </a> </li> </ul> </li> </ul> </nav> </devsite-footer-linkboxes> <devsite-footer-utility class="devsite-footer"> <div class="devsite-footer-utility nocontent"> <nav class="devsite-footer-utility-links" aria-label="Utility links"> <ul class="devsite-footer-utility-list"> <li class="devsite-footer-utility-item "> <a class="devsite-footer-utility-link gc-analytics-event" href="//about.google/" data-category="Site-Wide Custom Events" data-label="Footer About Google link" track-metadata-module="utility footer" track-metadata-position="footer" target="_blank" track-metadata-eventDetail="//about.google/" track-name="about google" track-type="footer link" > About Google </a> </li> <li class="devsite-footer-utility-item devsite-footer-privacy-link"> <a class="devsite-footer-utility-link gc-analytics-event" href="//policies.google.com/privacy" data-category="Site-Wide Custom Events" data-label="Footer Privacy link" track-metadata-module="utility footer" track-metadata-position="footer" track-type="footer link" track-metadata-eventDetail="//policies.google.com/privacy" track-name="privacy" target="_blank" > Privacy </a> </li> <li class="devsite-footer-utility-item "> <a class="devsite-footer-utility-link gc-analytics-event" href="//www.google.com/intl/en/policies/terms/regional.html" data-category="Site-Wide Custom Events" data-label="Footer Site terms link" target="_blank" track-metadata-module="utility footer" track-metadata-position="footer" track-type="footer link" track-metadata-eventDetail="//www.google.com/intl/en/policies/terms/regional.html" track-name="site terms" > Site terms </a> </li> <li class="devsite-footer-utility-item "> <a class="devsite-footer-utility-link gc-analytics-event" href="/product-terms/" data-category="Site-Wide Custom Events" data-label="Footer Google Cloud terms link" track-name="google cloud terms" track-metadata-position="footer" track-type="footer link" track-metadata-eventDetail="/product-terms/" track-metadata-module="utility footer" > Google Cloud terms </a> </li> <li class="devsite-footer-utility-item glue-cookie-notification-bar-control"> <a class="devsite-footer-utility-link gc-analytics-event" href="#" data-category="Site-Wide Custom Events" data-label="Footer Manage cookies link" track-metadata-module="utility footer" aria-hidden="true" track-type="footer link" track-name="Manage cookies" track-metadata-position="footer" track-metadata-eventDetail="#" > Manage cookies </a> </li> <li class="devsite-footer-utility-item devsite-footer-carbon-button"> <a class="devsite-footer-utility-link gc-analytics-event" href="/sustainability" data-category="Site-Wide Custom Events" data-label="Footer Our third decade of climate action: join us link" track-type="footer link" track-name="Our third decade of climate action: join us" track-metadata-position="footer" track-metadata-eventDetail="/sustainability/" track-metadata-module="utility footer" > Our third decade of climate action: join us </a> </li> <li class="devsite-footer-utility-item devsite-footer-utility-button"> <span class="devsite-footer-utility-description">Sign up for the Google Cloud newsletter</span> <a class="devsite-footer-utility-link gc-analytics-event" href="/newsletter/" data-category="Site-Wide Custom Events" data-label="Footer Subscribe link" track-metadata-eventDetail="/newsletter/" track-type="footer link" track-name="subscribe" track-metadata-module="utility footer" track-metadata-position="footer" > Subscribe </a> </li> </ul> <devsite-language-selector> <ul role="presentation"> <li role="presentation"> <a role="menuitem" lang="en" >English</a> </li> <li role="presentation"> <a role="menuitem" lang="de" >Deutsch</a> </li> <li role="presentation"> <a role="menuitem" lang="es_419" >Español – América Latina</a> </li> <li role="presentation"> <a role="menuitem" lang="fr" >Français</a> </li> <li role="presentation"> <a role="menuitem" lang="pt_br" >Português – Brasil</a> </li> <li role="presentation"> <a role="menuitem" lang="zh_cn" >中文 – 简体</a> </li> <li role="presentation"> <a role="menuitem" lang="ja" >日本語</a> </li> <li role="presentation"> <a role="menuitem" lang="ko" >한국어</a> </li> </ul> </devsite-language-selector> </nav> </div> </devsite-footer-utility> <devsite-panel></devsite-panel> </section></section> <devsite-sitemask></devsite-sitemask> <devsite-snackbar></devsite-snackbar> <devsite-tooltip ></devsite-tooltip> <devsite-heading-link></devsite-heading-link> <devsite-analytics> <script type="application/json" analytics>[]</script> <script type="application/json" tag-management>{"at": "True", "ga4": [], "ga4p": [], "gtm": [{"id": "GTM-5CVQBG", "purpose": 1}], "parameters": {"internalUser": "False", "language": {"machineTranslated": "False", "requested": "en", "served": "en"}, "pageType": "article", "projectName": "Vertex AI", "signedIn": "False", "tenant": "cloud", "recommendations": {"sourcePage": "", "sourceType": 0, "sourceRank": 0, "sourceIdenticalDescriptions": 0, "sourceTitleWords": 0, "sourceDescriptionWords": 0, "experiment": ""}, "experiment": {"ids": ""}}}</script> </devsite-analytics> <devsite-badger></devsite-badger> <cloudx-user></cloudx-user> <cloudx-free-trial-eligible-store freeTrialEligible='true'></cloudx-free-trial-eligible-store> <cloudx-pricing-socket></cloudx-pricing-socket> <cloudx-experiments type="TestAACodivertedExperiment" path="/virtual/TestAACodivertedExperiment/configureExperiment" location="SG" variant="variant2" ></cloudx-experiments> <cloudx-experiment-ids userCountry="SG" devsiteExperimentIdList="[39300012, 39300023, 39300118, 39300195, 39300241, 39300317, 39300320, 39300326, 39300346, 39300354, 39300364, 39300373, 39300412, 39300422, 39300436, 39300473, 39300488, 39300496, 39300498, 39300569]"> </cloudx-experiment-ids> <script nonce="YQ44WIyfKpnJh33yYeyN+qrOJk8SAd"> (function(d,e,v,s,i,t,E){d['GoogleDevelopersObject']=i; t=e.createElement(v);t.async=1;t.src=s;E=e.getElementsByTagName(v)[0]; E.parentNode.insertBefore(t,E);})(window, document, 'script', 'https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/js/app_loader.js', '[2,"en",null,"/js/devsite_app_module.js","https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f","https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud","https://cloud-dot-devsite-v2-prod.appspot.com",null,null,["/_pwa/cloud/manifest.json","https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/images/video-placeholder.svg","https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/favicons/onecloud/favicon.ico","https://www.gstatic.com/devrel-devsite/prod/v38a693baeb774512feb42f10aac8f755d8791ed41119b5be7a531f8e16f8279f/cloud/images/cloud-logo.svg","https://fonts.googleapis.com/css?family=Google+Sans:400,500,700|Google+Sans+Text:400,400italic,500,500italic,700,700italic|Roboto:400,400italic,500,500italic,700,700italic|Roboto+Mono:400,500,700&display=swap"],1,null,[1,6,8,12,14,17,21,25,50,52,63,70,75,76,80,87,91,92,93,97,98,100,101,102,103,104,105,107,108,109,110,112,113,116,117,118,120,122,124,125,126,127,129,130,131,132,133,134,135,136,138,140,141,147,148,149,151,152,156,157,158,159,161,163,164,168,169,170,179,180,182,183,186,191,193,196],"AIzaSyAP-jjEJBzmIyKR4F-3XITp8yM9T1gEEI8","AIzaSyB6xiKGDR5O3Ak2okS4rLkauxGUG7XP0hg","cloud.google.com","AIzaSyAQk0fBONSGUqCNznf6Krs82Ap1-NV6J4o","AIzaSyCCxcqdrZ_7QMeLCRY20bh_SXdAYqy70KY",null,null,null,["Cloud__enable_cloud_shell_fte_user_flow","Search__enable_suggestions_from_borg","Analytics__enable_clearcut_logging","Profiles__enable_page_saving","MiscFeatureFlags__developers_footer_dark_image","MiscFeatureFlags__emergency_css","EngEduTelemetry__enable_engedu_telemetry","Cloud__enable_cloudx_ping","Profiles__enable_join_program_group_endpoint","MiscFeatureFlags__enable_variable_operator","Search__enable_dynamic_content_confidential_banner","TpcFeatures__enable_mirror_tenant_redirects","CloudShell__cloud_shell_button","Search__scope_to_project_tenant","MiscFeatureFlags__enable_view_transitions","Profiles__enable_dashboard_curated_recommendations","Search__enable_ai_search_summaries","Cloud__enable_cloudx_experiment_ids","Profiles__enable_awarding_url","MiscFeatureFlags__enable_project_variables","CloudShell__cloud_code_overflow_menu","Cloud__enable_legacy_calculator_redirect","Profiles__enable_stripe_subscription_management","Search__enable_ai_search_summaries_restricted","Experiments__reqs_query_experiments","Search__enable_page_map","Profiles__enable_release_notes_notifications","BookNav__enable_tenant_cache_key","Profiles__enable_completecodelab_endpoint","Profiles__enable_public_developer_profiles","Cloud__enable_cloud_facet_chat","Profiles__require_profile_eligibility_for_signin","MiscFeatureFlags__developers_footer_image","Profiles__enable_recognition_badges","DevPro__enable_developer_subscriptions","Cloud__enable_free_trial_server_call","Profiles__enable_profile_collections","Concierge__enable_concierge_restricted","MiscFeatureFlags__enable_firebase_utm","Profiles__enable_completequiz_endpoint","Cloud__enable_cloud_dlp_service","Profiles__enable_developer_profiles_callout","Profiles__enable_complete_playlist_endpoint","Search__enable_ai_eligibility_checks","DevPro__enable_cloud_innovators_plus","Cloud__enable_llm_concierge_chat","Concierge__enable_pushui","MiscFeatureFlags__enable_explain_this_code","TpcFeatures__enable_unmirrored_page_left_nav","Cloud__enable_cloud_shell"],null,null,"AIzaSyBLEMok-5suZ67qRPzx0qUtbnLmyT_kCVE","https://developerscontentserving-pa.clients6.google.com","AIzaSyCM4QpTRSqP5qI4Dvjt4OAScIN8sOUlO-k","https://developerscontentsearch-pa.clients6.google.com",1,4,1,"https://developerprofiles-pa.clients6.google.com",[2,"cloud","Google Cloud","cloud.google.com",null,"cloud-dot-devsite-v2-prod.appspot.com",null,null,[1,1,null,null,null,null,null,null,null,null,null,[1],null,null,null,null,null,1,[1],[null,null,null,[1,20],"/terms/recommendations"],[1],null,[1],[1,null,1],[1,1,null,null,1,null,["/vertex-ai/"]]],null,[22,null,null,null,null,null,"/images/cloud-logo.svg","/images/favicons/onecloud/apple-icon.png",null,null,null,null,1,1,1,[6,5],[],null,null,[[],[],[],[],[],[],[],[]],null,1,null,null,null,null,[]],[],null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,null,[6,1,14,15,22,23,29,37],null,[[null,null,null,null,null,null,[1,[["docType","Choose a content type",[["ApiReference",null,null,null,null,null,null,null,null,"API reference"],["Sample",null,null,null,null,null,null,null,null,"Code sample"],["ReferenceArchitecture",null,null,null,null,null,null,null,null,"Reference architecture"],["Tutorial",null,null,null,null,null,null,null,null,"Tutorial"]]],["category","Choose a topic",[["AiAndMachineLearning",null,null,null,null,null,null,null,null,"Artificial intelligence and machine learning (AI/ML)"],["ApplicationDevelopment",null,null,null,null,null,null,null,null,"Application development"],["BigDataAndAnalytics",null,null,null,null,null,null,null,null,"Big data and analytics"],["Compute",null,null,null,null,null,null,null,null,"Compute"],["Containers",null,null,null,null,null,null,null,null,"Containers"],["Databases",null,null,null,null,null,null,null,null,"Databases"],["HybridCloud",null,null,null,null,null,null,null,null,"Hybrid and multicloud"],["LoggingAndMonitoring",null,null,null,null,null,null,null,null,"Logging and monitoring"],["Migrations",null,null,null,null,null,null,null,null,"Migrations"],["Networking",null,null,null,null,null,null,null,null,"Networking"],["SecurityAndCompliance",null,null,null,null,null,null,null,null,"Security and compliance"],["Serverless",null,null,null,null,null,null,null,null,"Serverless"],["Storage",null,null,null,null,null,null,null,null,"Storage"]]]]]],[1],null,1],[[null,null,null,null,null,["GTM-5CVQBG"],null,null,null,null,null,[["GTM-5CVQBG",2]],1],null,null,null,null,null,1],"mwETRvWii0eU5NUYprb0Y9z5GVbc",4,null,null,null,null,null,null,null,null,null,null,null,null,null,"cloud.devsite.google"],null,"pk_live_5170syrHvgGVmSx9sBrnWtA5luvk9BwnVcvIi7HizpwauFG96WedXsuXh790rtij9AmGllqPtMLfhe2RSwD6Pn38V00uBCydV4m",1]') </script> <devsite-a11y-announce></devsite-a11y-announce> </body> </html>

CINXE.COM

Vertex AI release notes | Google Cloud